Search Results

Search found 3710 results on 149 pages for 'databases'.

Page 41/149 | < Previous Page | 37 38 39 40 41 42 43 44 45 46 47 48  | Next Page >

  • Feature Updates to the Windows Azure Portal

    - by Clint Edmonson
    Lots of activity over at the Windows Azure portal this weekend, including some exciting new features and major improvements to existing features. Here are the highlights: Support for Managing Co-administrators Set up account co-administrators to allow others to share service management duties for each Azure subscription Import/Export support for SQL Databases Export existing SQL Azure databases to blob storage using SQL Server 2012’s BACPAC format. Create a new SQL Azure database from an existing BACPAC stored in blob storage Storage Container Management and Access Control Create blob storage containers directly within the portal Edit their public/private access settings Drill into storage containers and see the blobs contained within them Improved Cloud Service Status Notifications Detailed health status information about cloud services and roles as they transition between states Virtual Machine Experience Enhancements Option to automatically delete corresponding VHD files from blob storage when deleting VM disks Service Bus Management and Monitoring Ability to create and manage service bus Namespaces, Queues, Topics, Relays and Subscriptions Rich monitoring of Topics, Queues, and Subscriptions with detailed and customizable dashboard metrics Entity status (Topic, Queue, or Subscription) can be changed interactively via dashboard Direct links to the Access Control Services (ACS) namespaces when working with service bus access keys Media Services Monitoring Support Monitor encoding jobs that are queued for processing as well as active, failed and queued tasks for encoding jobs The above features are all now live in production and available to use immediately.  If you don’t already have a Windows Azure account, you can sign-up for a free trial and start using them today. Stay tuned to my twitter feed for Windows Azure announcements, updates, and links: @clinted Reference ID: P7VVJCM38V8R

    Read the article

  • How should I structure my database to gain maximum efficiently in this scenario?

    - by Bob Jansen
    I'm developing a PHP script that analyzes the web traffic of my clients websites. By placing a link to a javascript on the clients website (think of Google Analyses), my script harvests information like: the visitors IP address, reference link, current page link, user agent, etc. Now my clients can view these statistics via a control panel that I have build. These clients can also adjust profile settings, set firewall rules, create support tickets and pay invoices. Currently all the the traffic is stored in one table. You can imagine that this tabel would become very large as some my clients receive thousands of pageviews per day. Furthermore, all the traffic data of each client would be stored in the same table, creating a mess. This is the same for the firewall rules currently, and the invoice and support system. I'm looking for way to structure my database in a more organized way to hold large amounts of data of multiple users. This is the first project that I'm developing that deals with so much data, and would like to hear suggestions and tips. I was thinking of using multiple databases to structure the data. The main database will store users data (email,pass,id,etc) admin/website settings. Than each client will have an unique database labeled prefix_userid, which carry tables holding their traffic, invoice, and support ticket data. Would this be a solution, and would it slow down or speed up overall performances (that is spreading the data over muliple databases). I have a solid VPS, but would like to safe and be as effient as possible.

    Read the article

  • Parameterized Django models

    - by mgibsonbr
    In principle, a single Django application can be reused in two or more projects, providing functionality relevent to both. That implies that the same database structure (tables and relations) will be re-created identically in different databases, and most times this is not a problem (assuming the projects/databases are unrelated - for instance when someone downloads a complete app to use in their own projects). Sometimes, however, the models must be "tweaked" a little to better fit the problem needs. This can be accomplished by forking the app, but I wondered if there wouldn't be a better option in cases where the app designer can anticipate the most common customizations. For instance, if I have a model that could relate to another as one-to-one or one-to-many, I could specify the unique property as a parameter, that can be specified in the project's settings: class This(models.Model): other = models.ForeignKey(Other, unique=settings.OTHER_TO_THIS) Or if a model can relate to many others, I could create an intermediate table for each of them (thus enforcing referential integrity) instead of using generic fks: for related in settings.MODELS_RELATED_TO_OTHER: model_name = '%s_Other' % related globals()[model_name] = type(model_name, (models.Model,) { me:models.ForeignKey(find_model_class(related)), other:models.ForeignKey(Other), # Some other properties all intersection tables must have }) Etc. Let me stress out that I'm not proposing to change the models at runtime nor anything like that; once the parameters were defined and syncdb called for the first time, those parameters are not to be changed again (unless you're doing a schema migration). Is this a good design? Are there better ways to accomplish the same thing, or maybe drawbacks I coulnd't anticipate? This technique is meant to be used sparingly (only on apps meant to be reused in wildly different contexts, and only when a specific need of customization can be detected while the app model is being designed).

    Read the article

  • Centrally managing 100+ websites without bankrupting a small company

    - by palintropos
    I'm mainly interested in opinions on the trade-offs between having a single central server all the websites connect to as opposed to each website mirroring a subset of the master database with all the products in it. For example, will I run into severe performance issues (or even security issues, or restrictions) making queries to an offsite database? Will we hit scalability issues we can't handle early on from the sheer bandwidth required to maintain this? If we do go with something like a script that keeps smaller databases (each containing a subset of the central master data) in sync, what sorts of issues will we likely encounter there? I would really like the opinions of people far more knowledgeable than I am regarding the pros and cons of both setups and what headaches we are likely to encounter. CLARIFICATION: This should not be viewed as a question about whether we should implement one database vs multiple databases. This question has been answered numerous times. The question is regarding the pros and cons for a deployment like this having the ability to manage all the websites centrally (one server) vs trying to keep them all in sync if they each have their own db (multiple servers). REAL-WORLD EXAMPLE: We are a t-shirt company, and we have individual websites for our different kinds of t-shirts, but we're looking at a central order management integrated with our single shopping cart (which is ColdFusion + MySQL). Now, let's say we have a t-shirt that's on 10 of our websites and we change an image for it. Ideally we would change that in one place and the change would propagate, but how would we set this up?

    Read the article

  • Access a PLESK website before propagation?

    - by RCNeil
    My web host uses Plesk and I want to know if there is anyway to access and view a website (with PHP and other processes being functional) without propagation of the domain name? I have found countless forums on this but they are all pretty old (circa 01-04) and involve either tricking your localhost or SSH commands and some even result in terrible security risks. I would like to access a web page directory through a browser and see it's contents while having the PHP processes carry out... before I propagate it's potential domain name. People claim this is pointless but during a site migration why on earth would you not test a site before propagating it? I'm looking for something similar to what cPanel offers i.e. http://IP.ADDRESS./~mydomain.com The only solution I could think of is storing the site in a new directory of an already functional site and then setting up databases and testing the site once it's complete. Once tested and working I should be easily be able to migrate the files to the "new" domain name's root directory and just setup a new databases and then propagate the domain name. I can't believe that Plesk V10+ still does not have a site preview method that includes PHP, JS, and Flash ability.

    Read the article

  • How-To: Run CMSDK against a RAC cluster

    - by frank.closheim
    Using CMSDK in a production environment often requires a robust, reliable and failover enabled repository. When using Oracle Real Application Cluster (RAC) with your CMSDK repository you need to have a specific configuration in place to support such a setup. This post will explain the configuration steps required when running CMSDK 9.0.4.6 with Oracle WebLogic Server (WLS).In the previous CMSDK 9.0.4.2 version a RAC enabled connect string looked like this: (DESCRIPTION = (ADDRESS = (PROTOCOL = TCP)(HOST = rac1)(PORT = 1521))(ADDRESS = (PROTOCOL = TCP)(HOST = rac2)(PORT = 1521))(LOAD_BALANCE = NO)(FAILOVER = ON)(CONNECT_DATA =(SERVICE_NAME = rac)(failover_mode = (type=select)(method=basic)))CMSDK 9.0.4.6 makes use of data sources to connect to the underlying database. These data sources are configured inside your Application Server, such as Oracle WebLogic Server.In Oracle WebLogic Server 10.3.4, a single data source implementation has been introduced to support an RAC cluster. It responds to Fast Application Notification (FAN) events to provide Fast Connection Failover (FCF), Runtime Connection Load-Balancing (RCLB), and RAC instance graceful shutdown. XA affinity is supported at the global transaction Id level. The new feature is called WebLogic Active GridLink for RAC; which is implemented as the GridLink data source within WebLogic Server.This GridLink data source also works with Oracle Single Client Access Name (SCAN). SCAN is a feature used in RAC environments that provides a single name for clients to access any Oracle Database running in a cluster. You can think of SCAN as a cluster alias for databases in the cluster. The benefit is that the client’s connect information does not need to change if you add or remove nodes or databases in the cluster.The CMSDK 9.0.4.6 documentation describes how to create a regular JDBC data source named jdbc/OracleDS. Please refer to the following document which describes in detail how to create a GridLink data source in WLS.

    Read the article

  • Drupal migration failed

    - by Marco
    First of all, I'm new to Drupal and the work I have to do is some kind of too hard. My old colleague (webmaster) had a server with a multisite Drupal 6 installation. Sites and their dirs were (e.g.) Sites Site directory b.a.mycompany.com /drupal_install_dir/sites/b.a.mycompany.com c.a.mycompany.com /drupal_install_dir/sites/c.a.mycompany.com d.a.mycompany.com /drupal_install_dir/sites/d.a.mycompany.com Unluckily my colleague moved and server hdd aren't in my hands: all I have is a backup of /drupal_install_dir and three sql dumps (one for each site). I had to restore three sites, but changing them as z.mycompany.com/b z.mycompany.com/c z.mycompany.com/d Beeing a sysadmin, I Extracted tar.gz backup file under wwwroot (let's call full path to extracted directory /new_install_dir) Restored three databases Created mysql users and give them correct GRANTS on databases Then (trying to restore at least first site) I changed /new_install_dir/sites/settings.php putting correct database connection data and new basepath. But there is no way I can see my new site, simply it doesn't work. Watching /var/log/apache2/error.log I saw Drupal searching for main drupal database; so I created that db too setting user and grants, but dump file is empty. Well, now I can run something like install.php or update.php, but my site is not shown. Is there something I can do? Do I have to walk another way? Consider I searched the web, but I'm not able to find a guide that can help me for my problem. Ah, I forgot: before producing the backup, my colleague set site in maintenance mode. When I try to run z.mycompany.com/?q=user (trying to login) nothing happens. I'm really stuck...

    Read the article

  • How do I develop database-utilizing application in an agile/test-driven-development way?

    - by user39019
    I want to add databases (traditional client/server RDBMS's like Mysql/Postgresql as opposed to NoSQL, or embedded databases) to my toolbox as a developer. I've been using SQLite for simpler projects with only 1 client, but now I want to do more complicated things (ie, db-backed web development). I usually like following agile and/or test-driven-development principles. I generally code in Perl or Python. Questions: How do I test my code such that each run of the test suite starts with a 'pristine' state? Do I run a separate instance of the database server every test? Do I use a temporary database? How do I design my tables/schema so that it is flexible with respect to changing requirements? Do I start with an ORM for my language? Or do I stick to manually coding SQL? One thing I don't find appealing is having to change more than one thing (say, the CREATE TABLE statement and associated crud statements) for one change, b/c that's error prone. On the other hand, I expect ORM's to be a low slower and harder to debug than raw SQL. What is the general strategy for migrating data between one version of the program and a newer one? Do I carefully write ALTER TABLE statements between each version, or do I dump the data and import fresh in the new version?

    Read the article

  • SQL Server 2012 Express LocalDB &ndash; How to get started

    - by krislankford
    As many of you aware, SQL Server can be a bit of a pig when it comes to system resources on your development machine. As part of the 2012 products Microsoft has added SQL Server 2012 Express LocalDB which is a happy medium for myself when thinking about having to install a full blown SQL Server on my box. This however does not work in all cases for all development but if you are doing web or local client development then it should suffice. On the other hand, if you are working with technologies like SharePoint or trying to run Team Foundation Server on your local box then you will be out of luck while using LocalDB. To start of with, the localDB setup is delivered and packaged with Visual Studio 2012 RC. If you want to get the stand-alone installer you can download it here in either the 32 or 64 bit flavors. Once you get it installed you can start using it right away in either Visual Studio 2010 or the new Visual Studio 2012 RC. To get started you can open the SQL Server object explorer in Visual Studio by clicking   the menu option View –> SQL Server Object Explorer. This will bring up to the navigation pane where you can add a SQL Server. Once you add the SQL Server you will be prompted with the “Connect to Server” dialog to enter the server for which you can use “(localdb)\v11.0”. Click connect and you should be connected to your localDB where you can create and manage databases from Visual Studio 2010, Visual Studio 2012 or SSMS. Once you have started creating databases here you can use the database projects in Visual Studio with these database as well as use the (localdb)\v11.0 server name inside your connections string information for your development environment. Hope this helps someone get started with SQL Server 2012 Express LocalDB! It provides a great balance for developing against SQL Server 2012.

    Read the article

  • Postgres cannot connect to server

    - by user1408935
    Super stumped by why Postgres isn't working on a new app I just started. I've got it working for one app already. I'm using postgres.app, and it's running. I started a new app with rails new depot -d postgresql and then I went into the database.yml file and changed username to my $USER (which is what it is for the other app, which is working). So now my database.yml file has this development section: development: adapter: postgresql encoding: unicode database: depot_development pool: 5 username: <username> password: But when I run "rake db:create" or "rake db:create:all" I still got this error (in full, cause I don't know what's relevant): Couldn't create database for {"adapter"=>"postgresql", "encoding"=>"unicode", "database"=>"depot_development", "pool"=>5, "username"=>"<username>", "password"=>nil} could not connect to server: Permission denied Is the server running locally and accepting connections on Unix domain socket "/var/pgsql_socket/.s.PGSQL.5432"? /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/connection_adapters/postgresql_adapter.rb:1213:in `initialize' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/connection_adapters/postgresql_adapter.rb:1213:in `new' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/connection_adapters/postgresql_adapter.rb:1213:in `connect' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/connection_adapters/postgresql_adapter.rb:329:in `initialize' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/connection_adapters/postgresql_adapter.rb:28:in `new' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/connection_adapters/postgresql_adapter.rb:28:in `postgresql_connection' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/connection_adapters/abstract/connection_pool.rb:309:in `new_connection' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/connection_adapters/abstract/connection_pool.rb:319:in `checkout_new_connection' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/connection_adapters/abstract/connection_pool.rb:241:in `block (2 levels) in checkout' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/connection_adapters/abstract/connection_pool.rb:236:in `loop' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/connection_adapters/abstract/connection_pool.rb:236:in `block in checkout' /Users/<username>/.rvm/rubies/ruby-1.9.3-p194/lib/ruby/1.9.1/monitor.rb:211:in `mon_synchronize' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/connection_adapters/abstract/connection_pool.rb:233:in `checkout' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/connection_adapters/abstract/connection_pool.rb:96:in `block in connection' /Users/<username>/.rvm/rubies/ruby-1.9.3-p194/lib/ruby/1.9.1/monitor.rb:211:in `mon_synchronize' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/connection_adapters/abstract/connection_pool.rb:95:in `connection' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/connection_adapters/abstract/connection_pool.rb:404:in `retrieve_connection' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/connection_adapters/abstract/connection_specification.rb:170:in `retrieve_connection' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/connection_adapters/abstract/connection_specification.rb:144:in `connection' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/railties/databases.rake:107:in `rescue in create_database' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/railties/databases.rake:51:in `create_database' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/railties/databases.rake:40:in `block (3 levels) in <top (required)>' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/railties/databases.rake:40:in `each' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/gems/activerecord-3.2.8/lib/active_record/railties/databases.rake:40:in `block (2 levels) in <top (required)>' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/gems/rake-0.9.2.2/lib/rake/task.rb:205:in `call' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/gems/rake-0.9.2.2/lib/rake/task.rb:205:in `block in execute' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/gems/rake-0.9.2.2/lib/rake/task.rb:200:in `each' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/gems/rake-0.9.2.2/lib/rake/task.rb:200:in `execute' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/gems/rake-0.9.2.2/lib/rake/task.rb:158:in `block in invoke_with_call_chain' /Users/<username>/.rvm/rubies/ruby-1.9.3-p194/lib/ruby/1.9.1/monitor.rb:211:in `mon_synchronize' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/gems/rake-0.9.2.2/lib/rake/task.rb:151:in `invoke_with_call_chain' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/gems/rake-0.9.2.2/lib/rake/task.rb:144:in `invoke' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/gems/rake-0.9.2.2/lib/rake/application.rb:116:in `invoke_task' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/gems/rake-0.9.2.2/lib/rake/application.rb:94:in `block (2 levels) in top_level' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/gems/rake-0.9.2.2/lib/rake/application.rb:94:in `each' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/gems/rake-0.9.2.2/lib/rake/application.rb:94:in `block in top_level' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/gems/rake-0.9.2.2/lib/rake/application.rb:133:in `standard_exception_handling' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/gems/rake-0.9.2.2/lib/rake/application.rb:88:in `top_level' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/gems/rake-0.9.2.2/lib/rake/application.rb:66:in `block in run' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/gems/rake-0.9.2.2/lib/rake/application.rb:133:in `standard_exception_handling' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/gems/rake-0.9.2.2/lib/rake/application.rb:63:in `run' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/gems/rake-0.9.2.2/bin/rake:33:in `<top (required)>' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/bin/rake:19:in `load' /Users/<username>/.rvm/gems/ruby-1.9.3-p194@global/bin/rake:19:in `<main>' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/bin/ruby_noexec_wrapper:14:in `eval' /Users/<username>/.rvm/gems/ruby-1.9.3-p194/bin/ruby_noexec_wrapper:14:in `<main>' Couldn't create database for {"adapter"=>"postgresql", "encoding"=>"unicode", "database"=>"depot_test", "pool"=>5, "username"=>"<username>", "password"=>nil} I have tried createdb depot_development I have tried going into the psql environment and listing users (which included my username among them). In the same psql environment, I tried CREATE DATABASE depot; I've made sure that the pg gem is installed with bundle install, I've run "pg_ctl start", to which I got this response: pg_ctl: no database directory specified and environment variable PGDATA unset I ran "ps aux | grep postgres" to make sure postgres was running, to which I got this in return (which looks like it's doing OK, right?): <username> 10390 0.4 0.0 2425480 180 s000 R+ 6:15PM 0:00.00 grep postgres <username> 2907 0.0 0.0 2441604 464 ?? Ss 6:17PM 0:02.31 postgres: stats collector process <username> 2906 0.0 0.0 2445520 1664 ?? Ss 6:17PM 0:02.33 postgres: autovacuum launcher process <username> 2905 0.0 0.0 2445388 600 ?? Ss 6:17PM 0:09.25 postgres: wal writer process <username> 2904 0.0 0.0 2445388 1252 ?? Ss 6:17PM 0:12.08 postgres: writer process <username> 2902 0.0 0.0 2445388 3688 ?? S 6:17PM 0:00.54 /Applications/Postgres.app/Contents/MacOS/bin/postgres -D /Users/<username>/Library/Application Support/Postgres/var -p5432 The short of it, is I've been troubleshooting for a WHILE and have NO idea what's wrong. Any ideas? I'd really appreciate it, cause I'm pretty new to Rails, and this is a pretty disheartening roadblock. Thanks! EDIT -- Per request, posting the successful database.yml . It seems the difference is the inclusion of a password: development: adapter: postgresql encoding: unicode database: *******_development pool: 5 username: ******* password: ******* EDIT2 -- When I add a password to the .yml file, then run rake db:create again, I get this error. rake aborted! No Rakefile found (looking for: rakefile, Rakefile, rakefile.rb, Rakefile.rb)

    Read the article

  • Authoritative sources about Database vs. Flatfile decision

    - by FastAl
    <tldr>looking for a reference to a book or other undeniably authoritative source that gives reasons when you should choose a database vs. when you should choose other storage methods. I have provided an un-authoritative list of reasons about 2/3 of the way down this post.</tldr> I have a situation at my company where a database is being used where it would be better to use another solution (in this case, an auto-generated piece of source code that contains a static lookup table, searched by binary sort). Normally, a database would be an OK solution even though the problem does not require a database, e.g, none of the elements of ACID are needed, as it is read-only data, updated about every 3-5 years (also requiring other sourcecode changes), and fits in memory, and can be keyed into via binary search (a tad faster than db, but speed is not an issue). The problem is that this code runs on our enterprise server, but is shared with several PC platforms (some disconnected, some use a central DB, etc.), and parts of it are managed by multiple programming units, parts by the DBAs, parts even by mathematicians in another department, etc. These hit their own platform’s version of their databases (containing their own copy of the static data). What happens is that every implementation, every little change, something different goes wrong. There are many other issues as well. I can’t even use a flatfile, because one mode of running on our enterprise server does not have permission to read files (only databases, and of course, its own literal storage, e.g., in-source table). Of course, other parts of the system use databases in proper, less obscure manners; there is no problem with those parts. So why don’t we just change it? I don’t have administrative ability to force a change. But I’m affected because sometimes I have to help fix the problems, but mostly because it causes outages and tons of extra IT time by other programmers and d*mmit that makes me mad! The reason neither management, nor the designers of the system, can see the problem is that they propose a solution that won’t work: increase communication; implement more safeguards and standards; etc. But every time, in a different part of the already-pared-down but still multi-step processes, a few different diligent, hard-working, top performing IT personnel make a unique subtle error that causes it to fail, sometimes after the last round of testing! And in general these are not single-person failures, but understandable miscommunications. And communication at our company is actually better than most. People just don't think that's the case because they haven't dug into the matter. However, I have it on very good word from somebody with extensive formal study of sociology and psychology that the relatively small amount of less-than-proper database usage in this gigantic cross-platform multi-source, multi-language project is bureaucratically un-maintainable. Impossible. No chance. At least with Human Beings in the loop, and it can’t be automated. In addition, the management and developers who could change this, though intelligent and capable, don’t understand the rigidity of this ‘how humans are’ issue, and are not convincible on the matter. The reason putting the static data in sourcecode will solve the problem is, although the solution is less sexy than a database, it would function with no technical drawbacks; and since the sharing of sourcecode already works very well, you basically erase any database-related effort from this section of the project, along with all the drawbacks of it that are causing problems. OK, that’s the background, for the curious. I won’t be able to convince management that this is an unfixable sociological problem, and that the real solution is coding around these limits of human nature, just as you would code around a bug in a 3rd party component that you can’t change. So what I have to do is exploit the unsuitableness of the database solution, and not do it using logic, but rather authority. I am aware of many reasons, and posts on this site giving reasons for one over the other; I’m not looking for lists of reasons like these (although you can add a comment if I've miss a doozy): WHY USE A DATABASE? instead of flatfile/other DB vs. file: if you need... Random Read / Transparent search optimization Advanced / varied / customizable Searching and sorting capabilities Transaction/rollback Locks, semaphores Concurrency control / Shared users Security 1-many/m-m is easier Easy modification Scalability Load Balancing Random updates / inserts / deletes Advanced query Administrative control of design, etc. SQL / learning curve Debugging / Logging Centralized / Live Backup capabilities Cached queries / dvlp & cache execution plans Interleaved update/read Referential integrity, avoid redundant/missing/corrupt/out-of-sync data Reporting (from on olap or oltp db) / turnkey generation tools [Disadvantages:] Important to get right the first time - professional design - but only b/c it's meant to last s/w & h/w cost Usu. over a network, speed issue (best vs. best design vs. local=even then a separate process req's marshalling/netwk layers/inter-p comm) indicies and query processing can stand in the way of simple processing (vs. flatfile) WHY USE FLATFILE: If you only need... Sequential Row processing only Limited usage append only (no reading, no master key/update) Only Update the record you're reading (fixed length recs only) Too big to fit into memory If Local disk / read-ahead network connection Portability / small system Email / cut & Paste / store as document by novice - simple format Low design learning curve but high cost later WHY USE IN-MEMORY/TABLE (tables, arrays, etc.): if you need... Processing a single db/ff record that was imported Known size of data Static data if hardcoding the table Narrow, unchanging use (e.g., one program or proc) -includes a class that will be shared, but encapsulates its data manipulation Extreme speed needed / high transaction frequency Random access - but search is dependent on implementation Following are some other posts about the topic: http://stackoverflow.com/questions/1499239/database-vs-flat-text-file-what-are-some-technical-reasons-for-choosing-one-over http://stackoverflow.com/questions/332825/are-flat-file-databases-any-good http://stackoverflow.com/questions/2356851/database-vs-flat-files http://stackoverflow.com/questions/514455/databases-vs-plain-text/514530 What I’d like to know is if anybody could recommend a hard, authoritative source containing these reasons. I’m looking for a paper book I can buy, or a reputable website with whitepapers about the issue (e.g., Microsoft, IBM), not counting the user-generated content on those sites. This will have a greater change to elicit a change that I’m looking for: less wasted programmer time, and more reliable programs. Thanks very much for your help. You win a prize for reading such a large post!

    Read the article

  • How do I prove I should put a table of values in source code instead of a database table?

    - by FastAl
    <tldr>looking for a reference to a book or other undeniably authoritative source that gives reasons when you should choose a database vs. when you should choose other storage methods. I have provided an un-authoritative list of reasons about 2/3 of the way down this post.</tldr> I have a situation at my company where a database is being used where it would be better to use another solution (in this case, an auto-generated piece of source code that contains a static lookup table, searched by binary sort). Normally, a database would be an OK solution even though the problem does not require a database, e.g, none of the elements of ACID are needed, as it is read-only data, updated about every 3-5 years (also requiring other sourcecode changes), and fits in memory, and can be keyed into via binary search (a tad faster than db, but speed is not an issue). The problem is that this code runs on our enterprise server, but is shared with several PC platforms (some disconnected, some use a central DB, etc.), and parts of it are managed by multiple programming units, parts by the DBAs, parts even by mathematicians in another department, etc. These hit their own platform’s version of their databases (containing their own copy of the static data). What happens is that every implementation, every little change, something different goes wrong. There are many other issues as well. I can’t even use a flatfile, because one mode of running on our enterprise server does not have permission to read files (only databases, and of course, its own literal storage, e.g., in-source table). Of course, other parts of the system use databases in proper, less obscure manners; there is no problem with those parts. So why don’t we just change it? I don’t have administrative ability to force a change. But I’m affected because sometimes I have to help fix the problems, but mostly because it causes outages and tons of extra IT time by other programmers and d*mmit that makes me mad! The reason neither management, nor the designers of the system, can see the problem is that they propose a solution that won’t work: increase communication; implement more safeguards and standards; etc. But every time, in a different part of the already-pared-down but still multi-step processes, a few different diligent, hard-working, top performing IT personnel make a unique subtle error that causes it to fail, sometimes after the last round of testing! And in general these are not single-person failures, but understandable miscommunications. And communication at our company is actually better than most. People just don't think that's the case because they haven't dug into the matter. However, I have it on very good word from somebody with extensive formal study of sociology and psychology that the relatively small amount of less-than-proper database usage in this gigantic cross-platform multi-source, multi-language project is bureaucratically un-maintainable. Impossible. No chance. At least with Human Beings in the loop, and it can’t be automated. In addition, the management and developers who could change this, though intelligent and capable, don’t understand the rigidity of this ‘how humans are’ issue, and are not convincible on the matter. The reason putting the static data in sourcecode will solve the problem is, although the solution is less sexy than a database, it would function with no technical drawbacks; and since the sharing of sourcecode already works very well, you basically erase any database-related effort from this section of the project, along with all the drawbacks of it that are causing problems. OK, that’s the background, for the curious. I won’t be able to convince management that this is an unfixable sociological problem, and that the real solution is coding around these limits of human nature, just as you would code around a bug in a 3rd party component that you can’t change. So what I have to do is exploit the unsuitableness of the database solution, and not do it using logic, but rather authority. I am aware of many reasons, and posts on this site giving reasons for one over the other; I’m not looking for lists of reasons like these (although you can add a comment if I've miss a doozy): WHY USE A DATABASE? instead of flatfile/other DB vs. file: if you need... Random Read / Transparent search optimization Advanced / varied / customizable Searching and sorting capabilities Transaction/rollback Locks, semaphores Concurrency control / Shared users Security 1-many/m-m is easier Easy modification Scalability Load Balancing Random updates / inserts / deletes Advanced query Administrative control of design, etc. SQL / learning curve Debugging / Logging Centralized / Live Backup capabilities Cached queries / dvlp & cache execution plans Interleaved update/read Referential integrity, avoid redundant/missing/corrupt/out-of-sync data Reporting (from on olap or oltp db) / turnkey generation tools [Disadvantages:] Important to get right the first time - professional design - but only b/c it's meant to last s/w & h/w cost Usu. over a network, speed issue (best vs. best design vs. local=even then a separate process req's marshalling/netwk layers/inter-p comm) indicies and query processing can stand in the way of simple processing (vs. flatfile) WHY USE FLATFILE: If you only need... Sequential Row processing only Limited usage append only (no reading, no master key/update) Only Update the record you're reading (fixed length recs only) Too big to fit into memory If Local disk / read-ahead network connection Portability / small system Email / cut & Paste / store as document by novice - simple format Low design learning curve but high cost later WHY USE IN-MEMORY/TABLE (tables, arrays, etc.): if you need... Processing a single db/ff record that was imported Known size of data Static data if hardcoding the table Narrow, unchanging use (e.g., one program or proc) -includes a class that will be shared, but encapsulates its data manipulation Extreme speed needed / high transaction frequency Random access - but search is dependent on implementation Following are some other posts about the topic: http://stackoverflow.com/questions/1499239/database-vs-flat-text-file-what-are-some-technical-reasons-for-choosing-one-over http://stackoverflow.com/questions/332825/are-flat-file-databases-any-good http://stackoverflow.com/questions/2356851/database-vs-flat-files http://stackoverflow.com/questions/514455/databases-vs-plain-text/514530 What I’d like to know is if anybody could recommend a hard, authoritative source containing these reasons. I’m looking for a paper book I can buy, or a reputable website with whitepapers about the issue (e.g., Microsoft, IBM), not counting the user-generated content on those sites. This will have a greater change to elicit a change that I’m looking for: less wasted programmer time, and more reliable programs. Thanks very much for your help. You win a prize for reading such a large post!

    Read the article

  • Adopting DBVCS

    - by Wes McClure
    Identify early adopters Pick a small project with a small(ish) team.  This can be a legacy application or a green-field application. Strive to find a team of early adopters that will be eager to try something new. Get the team on board! Research Research the tool(s) that you want to use.  Some tools provide all of the features you would need while some only provide a slice of the pie.  DBVCS requires the ability to manage a set of change scripts that update a database from one version to the next.  Ideally a tool can track database versions and automatically apply updates.  The change script generation process can be manual, but having diff tools available to automatically generate it can really reduce the overhead to adoption.  Finally, an automated tool to generate a script file per database object is an added bonus as your version control system can quickly identify what was changed in a commit (add/del/modify), just like with code changes. Don’t settle on just one tool, identify several.  Then work with the team to evaluate the tools.  Have the team do some tests of the following scenarios with each tool: Baseline an existing database: can the migration tool work with legacy databases?  Caution: most migration platforms do not support baselines or have poor support, especially the fad of fluent APIs. Add/drop tables Add/drop procedures/functions/views Alter tables (rename columns, add columns, remove columns) Massage data – migrations sometimes involve changing data types that cannot be implicitly casted and require you to decide how the data is explicitly cast to the new type.  This is a requirement for a migrations platform.  Think about a case where you might want to combine fields, or move a field from one table to another, you wouldn’t want to lose the data. Run the tool via the command line.  If you cannot automate the tool in Continuous Integration what is the point? Create a copy of a database on demand. Backup/restore databases locally. Let the team give feedback and decide together, what tool they would like to try out. My recommendation at this point would be to include TSqlMigrations and RoundHouse as SQL based migration platforms.  In general I would recommend staying away from the fluent platforms as they often lack baseline capabilities and add overhead to learn a new API when SQL is already a very well known DSL.  Code migrations often get messy with procedures/views/functions as these have to be created with SQL and aren’t cross platform anyways.  IMO stick to SQL based migrations. Reconciling Production If your project is a legacy application, you will need to reconcile the current state of production with your development databases.  Find changes in production and bring them down to development, even if they are old and need to be removed.  Once complete, produce a baseline of either dev or prod as they are now in sync.  Commit this to your VCS of choice. Add whatever schema changes tracking mechanism your tool requires to your development database.  This often requires adding a table to track the schema version of that database.  Your tool should support doing this for you.  You can add this table to production when you do your next release. Script out any changes currently in dev.  Remove production artifacts that you brought down during reconciliation.  Add change scripts for any outstanding changes in dev since the last production release.  Commit these to your repository.   Say No to Shared Dev DBs Simply put, you wouldn’t dream of sharing a code checkout, why would you share a development database?  If you have a shared dev database, back it up, distribute the backups and take the shared version offline (including the dev db server once all projects are using DB VCS).  Doing DB VCS with a shared database is bound to cause problems as people won’t be able to easily script out their own changes from those that others are working on.   First prod release Copy prod to your beta/testing environment.  Add the schema changes table (or mechanism) and do a test run of your changes.  If successful you can schedule this to be run on production.   Evaluation After your first release, evaluate the pain points of the process.  Try to find tools or modifications to existing tools to help fix them.  Don’t leave stones unturned, iteratively evolve your tools and practices to make the process as seamless as possible.  This is why I suggest open source alternatives.  Nothing is set in stone, a good example was adding transactional support to TSqlMigrations.  We ran into situations where an update would break a database, so I added a feature to do transactional updates and rollback on errors!  Another good example is generating change scripts.  We have been manually making these for months now.  I found an open source project called Open DB Diff and integrated this with TSqlMigrations.  These were things we just accepted at the time when we began adopting our tool set.  Once we became comfortable with the base functionality, it was time to start automating more of the process.  Just like anything else with development, never be afraid to try to find tools to make your job easier!   Enjoy -Wes

    Read the article

  • SQL SERVER – SSMS: Backup and Restore Events Report

    - by Pinal Dave
    A DBA wears multiple hats and in fact does more than what an eye can see. One of the core task of a DBA is to take backups. This looks so trivial that most developers shrug this off as the only activity a DBA might be doing. I have huge respect for DBA’s all around the world because even if they seem cool with all the scripting, automation, maintenance works round the clock to keep the business working almost 365 days 24×7, their worth is knowing that one day when the systems / HDD crashes and you have an important delivery to make. So these backup tasks / maintenance jobs that have been done come handy and are no more trivial as they might seem to be as considered by many. So the important question like: “When was the last backup taken?”, “How much time did the last backup take?”, “What type of backup was taken last?” etc are tricky questions and this report lands answers to the same in a jiffy. So the SSMS report, we are talking can be used to find backups and restore operation done for the selected database. Whenever we perform any backup or restore operation, the information is stored in the msdb database. This report can utilize that information and provide information about the size, time taken and also the file location for those operations. Here is how this report can be launched.   Once we launch this report, we can see 4 major sections shown as listed below. Average Time Taken For Backup Operations Successful Backup Operations Backup Operation Errors Successful Restore Operations Let us look at each section next. Average Time Taken For Backup Operations Information shown in “Average Time Taken For Backup Operations” section is taken from a backupset table in the msdb database. Here is the query and the expanded version of that particular section USE msdb; SELECT (ROW_NUMBER() OVER (ORDER BY t1.TYPE))%2 AS l1 ,       1 AS l2 ,       1 AS l3 ,       t1.TYPE AS [type] ,       (AVG(DATEDIFF(ss,backup_start_date, backup_finish_date)))/60.0 AS AverageBackupDuration FROM backupset t1 INNER JOIN sys.databases t3 ON ( t1.database_name = t3.name) WHERE t3.name = N'AdventureWorks2014' GROUP BY t1.TYPE ORDER BY t1.TYPE On my small database the time taken for differential backup was less than a minute, hence the value of zero is displayed. This is an important piece of backup operation which might help you in planning maintenance windows. Successful Backup Operations Here is the expanded version of this section.   This information is derived from various backup tracking tables from msdb database.  Here is the simplified version of the query which can be used separately as well. SELECT * FROM sys.databases t1 INNER JOIN backupset t3 ON (t3.database_name = t1.name) LEFT OUTER JOIN backupmediaset t5 ON ( t3.media_set_id = t5.media_set_id) LEFT OUTER JOIN backupmediafamily t6 ON ( t6.media_set_id = t5.media_set_id) WHERE (t1.name = N'AdventureWorks2014') ORDER BY backup_start_date DESC,t3.backup_set_id,t6.physical_device_name; The report does some calculations to show the data in a more readable format. For example, the backup size is shown in KB, MB or GB. I have expanded first row by clicking on (+) on “Device type” column. That has shown me the path of the physical backup file. Personally looking at this section, the Backup Size, Device Type and Backup Name are critical and are worth a note. As mentioned in the previous section, this section also has the Duration embedded inside it. Backup Operation Errors This section of the report gets data from default trace. You might wonder how. One of the event which is tracked by default trace is “ErrorLog”. This means that whatever message is written to errorlog gets written to default trace file as well. Interestingly, whenever there is a backup failure, an error message is written to ERRORLOG and hence default trace. This section takes advantage of that and shows the information. We can read below message under this section, which confirms above logic. No backup operations errors occurred for (AdventureWorks2014) database in the recent past or default trace is not enabled. Successful Restore Operations This section may not be very useful in production server (do you perform a restore of database?) but might be useful in the development and log shipping secondary environment, where we might be interested to see restore operations for a particular database. Here is the expanded version of the section. To fill this section of the report, I have restored the same backups which were taken to populate earlier sections. Here is the simplified version of the query used to populate this output. USE msdb; SELECT * FROM restorehistory t1 LEFT OUTER JOIN restorefile t2 ON ( t1.restore_history_id = t2.restore_history_id) LEFT OUTER JOIN backupset t3 ON ( t1.backup_set_id = t3.backup_set_id) WHERE t1.destination_database_name = N'AdventureWorks2014' ORDER BY restore_date DESC,  t1.restore_history_id,t2.destination_phys_name Have you ever looked at the backup strategy of your key databases? Are they in sync and do we have scope for improvements? Then this is the report to analyze after a week or month of maintenance plans running in your database. Do chime in with what are the strategies you are using in your environments. Reference: Pinal Dave (http://blog.sqlauthority.com)Filed under: PostADay, SQL, SQL Authority, SQL Backup and Restore, SQL Query, SQL Server, SQL Server Management Studio, SQL Tips and Tricks, T SQL Tagged: SQL Reports

    Read the article

  • How to restore your production database without needing additional storage

    - by David Atkinson
    Production databases can get very large. This in itself is to be expected, but when a copy of the database is needed the database must be restored, requiring additional and costly storage.  For example, if you want to give each developer a full copy of your production server, you’ll need n times the storage cost for your n-developer team. The same is true for any test databases that are created during the course of your project lifecycle. If you’ve read my previous blog posts, you’ll be aware that I’ve been focusing on the database continuous integration theme. In my CI setup I create a “production”-equivalent database directly from its source control representation, and use this to test my upgrade scripts. Despite this being a perfectly valid and practical thing to do as part of a CI setup, it’s not the exact equivalent to running the upgrade script on a copy of the actual production database. So why shouldn’t I instead simply restore the most recent production backup as part of my CI process? There are two reasons why this would be impractical. 1. My CI environment isn’t an exact copy of my production environment. Indeed, this would be the case in a perfect world, and it is strongly recommended as a good practice if you follow Jez Humble and David Farley’s “Continuous Delivery” teachings, but in practical terms this might not always be possible, especially where storage is concerned. It may just not be possible to restore a huge production database on the environment you’ve been allotted. 2. It’s not just about the storage requirements, it’s also the time it takes to do the restore. The whole point of continuous integration is that you are alerted as early as possible whether the build (yes, the database upgrade script counts!) is broken. If I have to run an hour-long restore each time I commit a change to source control I’m just not going to get the feedback quickly enough to react. So what’s the solution? Red Gate has a technology, SQL Virtual Restore, that is able to restore a database without using up additional storage. Although this sounds too good to be true, the explanation is quite simple (although I’m sure the technical implementation details under the hood are quite complex!) Instead of restoring the backup in the conventional sense, SQL Virtual Restore will effectively mount the backup using its HyperBac technology. It creates a data and log file, .vmdf, and .vldf, that becomes the delta between the .bak file and the virtual database. This means that both read and write operations are permitted on a virtual database as from SQL Server’s point of view it is no different from a conventional database. Instead of doubling the storage requirements upon a restore, there is no ‘duplicate’ storage requirements, other than the trivially small virtual log and data files (see illustration below). The benefit is magnified the more databases you mount to the same backup file. This technique could be used to provide a large development team a full development instance of a large production database. It is also incredibly easy to set up. Once SQL Virtual Restore is installed, you simply run a conventional RESTORE command to create the virtual database. This is what I have running as part of a nightly “release test” process triggered by my CI tool. RESTORE DATABASE WidgetProduction_Virtual FROM DISK=N'D:\VirtualDatabase\WidgetProduction.bak' WITH MOVE N'WidgetProduction' TO N'C:\WidgetWF\ProdBackup\WidgetProduction_WidgetProduction_Virtual.vmdf', MOVE N'WidgetProduction_log' TO N'C:\WidgetWF\ProdBackup\WidgetProduction_log_WidgetProduction_Virtual.vldf', NORECOVERY, STATS=1, REPLACE GO RESTORE DATABASE WidgetProduction_Virtual WITH RECOVERY   Note the only change from what you would do normally is the naming of the .vmdf and .vldf files. SQL Virtual Restore intercepts this by monitoring the extension and applies its magic, ensuring the ‘virtual’ restore happens rather than the conventional storage-heavy restore. My automated release test then applies the upgrade scripts to the virtual production database and runs some validation tests, giving me confidence that were I to run this on production for real, all would go smoothly. For illustration, here is my 8Gb production database: And its corresponding backup file: Here are the .vldf and .vmdf files, which represent the only additional used storage for the new database following the virtual restore.   The beauty of this product is its simplicity. Once it is installed, the interaction with the backup and virtual database is exactly the same as before, as the clever stuff is being done at a lower level. SQL Virtual Restore can be downloaded as a fully functional 14-day trial. Technorati Tags: SQL Server

    Read the article

  • How to restore your production database without needing additional storage

    - by David Atkinson
    Production databases can get very large. This in itself is to be expected, but when a copy of the database is needed the database must be restored, requiring additional and costly storage.  For example, if you want to give each developer a full copy of your production server, you'll need n times the storage cost for your n-developer team. The same is true for any test databases that are created during the course of your project lifecycle. If you've read my previous blog posts, you'll be aware that I've been focusing on the database continuous integration theme. In my CI setup I create a "production"-equivalent database directly from its source control representation, and use this to test my upgrade scripts. Despite this being a perfectly valid and practical thing to do as part of a CI setup, it's not the exact equivalent to running the upgrade script on a copy of the actual production database. So why shouldn't I instead simply restore the most recent production backup as part of my CI process? There are two reasons why this would be impractical. 1. My CI environment isn't an exact copy of my production environment. Indeed, this would be the case in a perfect world, and it is strongly recommended as a good practice if you follow Jez Humble and David Farley's "Continuous Delivery" teachings, but in practical terms this might not always be possible, especially where storage is concerned. It may just not be possible to restore a huge production database on the environment you've been allotted. 2. It's not just about the storage requirements, it's also the time it takes to do the restore. The whole point of continuous integration is that you are alerted as early as possible whether the build (yes, the database upgrade script counts!) is broken. If I have to run an hour-long restore each time I commit a change to source control I'm just not going to get the feedback quickly enough to react. So what's the solution? Red Gate has a technology, SQL Virtual Restore, that is able to restore a database without using up additional storage. Although this sounds too good to be true, the explanation is quite simple (although I'm sure the technical implementation details under the hood are quite complex!) Instead of restoring the backup in the conventional sense, SQL Virtual Restore will effectively mount the backup using its HyperBac technology. It creates a data and log file, .vmdf, and .vldf, that becomes the delta between the .bak file and the virtual database. This means that both read and write operations are permitted on a virtual database as from SQL Server's point of view it is no different from a conventional database. Instead of doubling the storage requirements upon a restore, there is no 'duplicate' storage requirements, other than the trivially small virtual log and data files (see illustration below). The benefit is magnified the more databases you mount to the same backup file. This technique could be used to provide a large development team a full development instance of a large production database. It is also incredibly easy to set up. Once SQL Virtual Restore is installed, you simply run a conventional RESTORE command to create the virtual database. This is what I have running as part of a nightly "release test" process triggered by my CI tool. RESTORE DATABASE WidgetProduction_virtual FROM DISK=N'C:\WidgetWF\ProdBackup\WidgetProduction.bak' WITH MOVE N'WidgetProduction' TO N'C:\WidgetWF\ProdBackup\WidgetProduction_WidgetProduction_Virtual.vmdf', MOVE N'WidgetProduction_log' TO N'C:\WidgetWF\ProdBackup\WidgetProduction_log_WidgetProduction_Virtual.vldf', NORECOVERY, STATS=1, REPLACE GO RESTORE DATABASE mydatabase WITH RECOVERY   Note the only change from what you would do normally is the naming of the .vmdf and .vldf files. SQL Virtual Restore intercepts this by monitoring the extension and applies its magic, ensuring the 'virtual' restore happens rather than the conventional storage-heavy restore. My automated release test then applies the upgrade scripts to the virtual production database and runs some validation tests, giving me confidence that were I to run this on production for real, all would go smoothly. For illustration, here is my 8Gb production database: And its corresponding backup file: Here are the .vldf and .vmdf files, which represent the only additional used storage for the new database following the virtual restore.   The beauty of this product is its simplicity. Once it is installed, the interaction with the backup and virtual database is exactly the same as before, as the clever stuff is being done at a lower level. SQL Virtual Restore can be downloaded as a fully functional 14-day trial. Technorati Tags: SQL Server

    Read the article

  • How to Answer a Stupid Interview Question the Right Way

    - by AjarnMark
    Have you ever been asked a stupid question during an interview; one that seemed to have no relation to the job responsibilities at all?  Tech people are often caught off-guard by these apparently irrelevant questions, but there is a way you can turn these to your favor.  Here is one idea. While chatting with a couple of folks between sessions at SQLSaturday 43 last weekend, one of them expressed frustration over a seemingly ridiculous and trivial question that she was asked during an interview, and she believes it cost her the job opportunity.  The question, as I remember it being described was, “What is the largest byte measurement?”.  The candidate made up a guess (“zetabyte”) during the interview, which is actually closer than she may have realized.  According to Wikipedia, there is a measurement known as zettabyte which is 10^21, and the largest one listed there is yottabyte at 10^24. My first reaction to this question was, “That’s just a hiring manager that doesn’t really know what they’re looking for in a candidate.  Furthermore, this tells me that this manager really does not understand how to build a team.”  In most companies, team interaction is more important than uber-knowledge.  I didn’t ask, but this could also be another geek on the team trying to establish their Alpha-Geek stature.  I suppose that there are a few, very few, companies that can build their businesses on hiring only the extreme alpha-geeks, but that certainly does not represent the majority of businesses in America. My friend who was there suggested that the appropriate response to this silly question would be, “And how does this apply to the work I will be doing?” Of course this is an understandable response when you’re frustrated because you know you can handle the technical aspects of the job, and it seems like the interviewer is just being silly.  But it is also a direct challenge, which may not be the best approach in interviewing.  I do have to admit, though, that there are those folks who just won’t respect you until you do challenge them, but again, I don’t think that is the majority. So after some thought, here is my suggestion: “Well, I know that there are petabytes and exabytes and things even larger than that, but I haven’t been keeping up on my list of Greek prefixes that have not yet been used, so I would have to look up the exact answer if you need it.  However, I have worked with databases as large as 30 Terabytes.  How big are the largest databases here at X Corporation?”  Perhaps with a follow-up of, “Typically, what I have seen in companies that have databases of your size, is that the three biggest challenges they face are: A, B, and C.  What would you say are the top 3 concerns that you would like the person you hire to be able to address?…Here is how I have dealt with those concerns in the past (or ‘Here is how I would tackle those issues for you…’).” Wait! What just happened?!  We took a seemingly irrelevant and frustrating question and turned it around into an opportunity to highlight our relevant skills and guide the conversation back in a direction more to our liking and benefit.  In more generic terms, here is what we did: Admit that you don’t know the specific answer off the top of your head, but can get it if it’s truly important to the company. Maybe for some reason it really is important to them. Mention something similar or related that you do know, reassuring them that you do have some knowledge in that subject area. Draw a parallel to your past work experience. Ask follow-up questions about the company’s specific needs and discuss how you can fulfill those. This type of thing requires practice and some forethought.  I didn’t come up with this answer until a day later, which is too late when you’re interviewing.  I still think it is silly for an interviewer to ask something like that, but at least this is one way to spin it to your advantage while you consider whether you really want to work for someone who would ask a thing like that.  Remember, interviewing is a two-way process.  You’re deciding whether you want to work there just as much as they are deciding whether they want you. There is always the possibility that this was a calculated maneuver on the part of the hiring manager just to see how quickly you think on your feet and how you handle stupid questions.  Maybe he knows something about the work environment and he’s trying to gauge whether you’ll actually fit in okay.  And if that’s the case, then the above response still works quite well.

    Read the article

  • How to restore your production database without needing additional storage

    - by David Atkinson
    Production databases can get very large. This in itself is to be expected, but when a copy of the database is needed the database must be restored, requiring additional and costly storage.  For example, if you want to give each developer a full copy of your production server, you'll need n times the storage cost for your n-developer team. The same is true for any test databases that are created during the course of your project lifecycle. If you've read my previous blog posts, you'll be aware that I've been focusing on the database continuous integration theme. In my CI setup I create a "production"-equivalent database directly from its source control representation, and use this to test my upgrade scripts. Despite this being a perfectly valid and practical thing to do as part of a CI setup, it's not the exact equivalent to running the upgrade script on a copy of the actual production database. So why shouldn't I instead simply restore the most recent production backup as part of my CI process? There are two reasons why this would be impractical. 1. My CI environment isn't an exact copy of my production environment. Indeed, this would be the case in a perfect world, and it is strongly recommended as a good practice if you follow Jez Humble and David Farley's "Continuous Delivery" teachings, but in practical terms this might not always be possible, especially where storage is concerned. It may just not be possible to restore a huge production database on the environment you've been allotted. 2. It's not just about the storage requirements, it's also the time it takes to do the restore. The whole point of continuous integration is that you are alerted as early as possible whether the build (yes, the database upgrade script counts!) is broken. If I have to run an hour-long restore each time I commit a change to source control I'm just not going to get the feedback quickly enough to react. So what's the solution? Red Gate has a technology, SQL Virtual Restore, that is able to restore a database without using up additional storage. Although this sounds too good to be true, the explanation is quite simple (although I'm sure the technical implementation details under the hood are quite complex!) Instead of restoring the backup in the conventional sense, SQL Virtual Restore will effectively mount the backup using its HyperBac technology. It creates a data and log file, .vmdf, and .vldf, that becomes the delta between the .bak file and the virtual database. This means that both read and write operations are permitted on a virtual database as from SQL Server's point of view it is no different from a conventional database. Instead of doubling the storage requirements upon a restore, there is no 'duplicate' storage requirements, other than the trivially small virtual log and data files (see illustration below). The benefit is magnified the more databases you mount to the same backup file. This technique could be used to provide a large development team a full development instance of a large production database. It is also incredibly easy to set up. Once SQL Virtual Restore is installed, you simply run a conventional RESTORE command to create the virtual database. This is what I have running as part of a nightly "release test" process triggered by my CI tool. RESTORE DATABASE WidgetProduction_virtual FROM DISK=N'C:\WidgetWF\ProdBackup\WidgetProduction.bak' WITH MOVE N'WidgetProduction' TO N'C:\WidgetWF\ProdBackup\WidgetProduction_WidgetProduction_Virtual.vmdf', MOVE N'WidgetProduction_log' TO N'C:\WidgetWF\ProdBackup\WidgetProduction_log_WidgetProduction_Virtual.vldf', NORECOVERY, STATS=1, REPLACE GO RESTORE DATABASE mydatabase WITH RECOVERY   Note the only change from what you would do normally is the naming of the .vmdf and .vldf files. SQL Virtual Restore intercepts this by monitoring the extension and applies its magic, ensuring the 'virtual' restore happens rather than the conventional storage-heavy restore. My automated release test then applies the upgrade scripts to the virtual production database and runs some validation tests, giving me confidence that were I to run this on production for real, all would go smoothly. For illustration, here is my 8Gb production database: And its corresponding backup file: Here are the .vldf and .vmdf files, which represent the only additional used storage for the new database following the virtual restore.   The beauty of this product is its simplicity. Once it is installed, the interaction with the backup and virtual database is exactly the same as before, as the clever stuff is being done at a lower level. SQL Virtual Restore can be downloaded as a fully functional 14-day trial. Technorati Tags: SQL Server

    Read the article

  • Interview with Geoff Bones, developer on SQL Storage Compress

    - by red(at)work
    How did you come to be working at Red Gate? I've been working at Red Gate for nine months; before that I had been at a multinational engineering company. A number of my colleagues had left to work at Red Gate and spoke very highly of it, but I was happy in my role and thought, 'It can't be that great there, surely? They'll be back!' Then one day I visited to catch up them over lunch in the Red Gate canteen. I was so impressed with what I found there, that, three days later, I'd applied for a role as a developer. And how did you get into software development? My first job out of university was working as a systems programmer on IBM mainframes. This was quite a while ago: there was a lot of assembler and loading programs from tape drives and that kind of stuff. I learned a lot about how computers work, and this stood me in good stead when I moved over the development in the 90s. What's the best thing about working as a developer at Red Gate? Where should I start? One of the great things as a developer at Red Gate is the useful feedback and close contact we have with the people who use our products, either directly at trade shows and other events or through information coming through the product managers. The company's whole ethos is built around assisting the user, and this is in big contrast to my previous development roles. We aim to produce tools that people really want to use, that they enjoy using, and, as a developer, this is a great thing to aim for and a great feeling when we get it right. At Red Gate we also try to cut out the things that distract and stop us doing our jobs. As a developer, this means that I can focus on the code and the product I'm working on, knowing that others are doing a first-class job of making sure that the builds are running smoothly and that I'm getting great feedback from the testers. We keep our process light and effective, as we want to produce great software more than we want to produce great audit trails. Tell us a bit about the products you are currently working on. You mean HyperBac? First let me explain a bit about what HyperBac is. At heart it's a compression and encryption technology, but with a few added features that open up a wealth of really exciting possibilities. Right now we have the HyperBac technology in just three products: SQL HyperBac, SQL Virtual Restore and SQL Storage Compress, but we're only starting to develop what it can do. My personal favourite is SQL Virtual Restore; for example, I love the way you can use it to run independent test databases that are all backed by a single compressed backup. I don't think the market yet realises the kind of things you do once you are using these products. On the other hand, the benefits of SQL Storage Compress are straightforward: run your databases but use only 20% of the disk space. Databases are getting larger and larger, and, as they do, so does your ROI. What's a typical day for you? My days are pretty varied. We have our daily team stand-up meeting and then sometimes I will work alone on a current issue, or I'll be pair programming with one of my colleagues. From time to time we give half a day up to future planning with the team, when we look at the long and short term aims for the product and working out the development priorities. I also get to go to conferences and events, which is unusual for a development role and gives me the chance to meet and talk to our customers directly. Have you noticed anything different about developing tools for DBAs rather than other IT kinds of user? It seems to me that DBAs are quite independent minded; they know exactly what the problem they are facing is, and often have a solution in mind before they begin to look for what's on the market. This means that they're likely to cherry-pick tools from a range of vendors, picking the ones that are the best fit for them and that disrupt their environments the least. When I've met with DBAs, I've often been very impressed at their ability to summarise their set up, the issues, the obstacles they face when implementing a tool and their plans for their environment. It's easier to develop products for this audience as they give such a detailed overview of their needs, and I feel I understand their problems.

    Read the article

  • Disaster Recovery Discovery

    - by Rodney Landrum
    Last weekend I joined several of my IT staff on a mission to perform a DR test in our remote CoLo center in a large South East city of the US. Can I be more obtuse? The goal was simple for me as the sole DBA in a throng of Windows, Storage, Network and SAN admins – restore the databases and make them work. There were 4 applications that back ended to 7 SQL Server databases on 4 different SQL Server instances. We would maintain the original server names, but beyond that it was fair game. We had time to prepare so I was able to script out or otherwise automate the recovery process. I used sp_help_revlogin for three of the servers, a bit of a cheat actually because restoring the Master database on the target DR servers was the specified course of action according to the DR procedures ( the caveat “IF REQUIRED” left it open to interpretation. I really wanted to avoid the step of restoring Master for a number of reasons but mainly because I did not want to deal with issues starting SQL Services afterward. Having to account for the location of TempDB and the version conflicts of the resource DBs were just two of the battles I chose not to fight. Not to mention other system database location problems that might arise and prevent SQL from starting.  I was going to have to restore all of the user databases anyway, so I would not really gain any benefit, outside of logins, for taking the time to restore the source Master database over the newly installed one on the fresh server. What I wanted was the ability to restore the Master database as a user database, call it Master_Mine, from a backup on the source system and then use that restored database to script the SQL Logins and passwords on the DR systems. While I did not attempt this on the trip, the thought stuck in my mind and this past week I succeeded at scripting user accounts and passwords using only a restored copy of the Master database. Granted there were several challenges to overcome.  Also, as is usual for any work like this the usual disclaimers apply:  This is not something that I would imagine Microsoft would condone or support and this was really only an experiment for me to learn if it was even possible. While I have tested the process with success, I do not know that I would use this technique in a documented procedure because future updates for SQL Server will render this technique non-functional. I thought at first, incorrectly of course, that I could use sp_help_revlogin on a restored copy of the master database I named Master_Mine.   Since sp_help_revlogin uses system schema objects, sys.syslogins and sys.server_principals, this was not going to work because all results would come from the main Master database. To test this I added a SQL login via SSMS, backed up Master, restored  it as Master_Mine, and then deleted the login.  Even though the test account I created should presumably still be in the Master_Mine database, I should be able to get to it and script out its creation with its password hash so that I would not need to know the password, but any applications that stored that password would not have to be altered in the DR scenario. They would just work as expected. Once I realized that would not work I began looking deeper.  Knowing that sys.syslogins and sys.server_principals are system views, their underlying code should be available with sp_helptext, right? They were. And this led me to discover the two tables sys.sysxlgns and sys.sysprivs, where the data I needed was stored. These tables existed in both the real Master and the restored copy, Master_Mine.  I used this information to tweak the sp_help_revlogin stored procedure to use these tables instead to create the logins cursor used in sp_help_revlogin. For the password hash,  sp_help_revlogin uses the function LoginProperty() which takes a user name and option ‘passwordhash’ to return the hash for the user. Unfortunately, it requires the login to exist in the Master database. This would not work. So another slight modification I had to make was to pull the password hash itself (pwdhash from sys.sysxlgns) into the logins cursor and comment out the section of sp_help_revlogin that uses LoginProperty. Instead, I pass the pwdhash value as the variable @PWD_varbinary to the sp_hexadecimal stored procedure which is also created by and used within the code provided by Microsoft in the link above for sp_help_revlogin. The final challenge: sys.sysxlgns and sys.server_principals are visible only within a Dedicated Administrator Connection (DAC) query window in SSMS or within SQLCDMD.  To open a DAC connection you have to be logged in on the SQL Server itself, via RDP in my case,  and you preface the server name in the query connection with ADMIN:, so that the server connection looks like ADMIN:ServerName. From there you can create the modified stored procedure in the restored copy of a Master database from a source system as whatever name you like, and then run the modified stored procedure. I named my new stored procedure usp_help_revlogin_MyMaster. Upon execution I was happy to see the logins and password hashes that I needed to apply from the source Master database without having to restore over the new Master system database and without the need to access the original server (assuming it was down due to whatever disaster put it in that state). You will note that I am not providing full code samples here of the modifications. I will say that it was a slight bit of work and anyone who needed to do this for whatever reason, could fairly easily roll their own solution with the information provided herein.  My goal, as I said was to prove that this could be done and provide another option if required to ease the burden of getting SQL Servers up and available in an emergency situation where alternatives may be more challenging or otherwise unavailable.  

    Read the article

  • Differences Between NHibernate and Entity Framework

    - by Ricardo Peres
    Introduction NHibernate and Entity Framework are two of the most popular O/RM frameworks on the .NET world. Although they share some functionality, there are some aspects on which they are quite different. This post will describe this differences and will hopefully help you get started with the one you know less. Mind you, this is a personal selection of features to compare, it is by no way an exhaustive list. History First, a bit of history. NHibernate is an open-source project that was first ported from Java’s venerable Hibernate framework, one of the first O/RM frameworks, but nowadays it is not tied to it, for example, it has .NET specific features, and has evolved in different ways from those of its Java counterpart. Current version is 3.3, with 3.4 on the horizon. It currently targets .NET 3.5, but can be used as well in .NET 4, it only makes no use of any of its specific functionality. You can find its home page at NHForge. Entity Framework 1 came out with .NET 3.5 and is now on its second major version, despite being version 4. Code First sits on top of it and but came separately and will also continue to be released out of line with major .NET distributions. It is currently on version 4.3.1 and version 5 will be released together with .NET Framework 4.5. All versions will target the current version of .NET, at the time of their release. Its home location is located at MSDN. Architecture In NHibernate, there is a separation between the Unit of Work and the configuration and model instances. You start off by creating a Configuration object, where you specify all global NHibernate settings such as the database and dialect to use, the batch sizes, the mappings, etc, then you build an ISessionFactory from it. The ISessionFactory holds model and metadata that is tied to a particular database and to the settings that came from the Configuration object, and, there will typically be only one instance of each in a process. Finally, you create instances of ISession from the ISessionFactory, which is the NHibernate representation of the Unit of Work and Identity Map. This is a lightweight object, it basically opens and closes a database connection as required and keeps track of the entities associated with it. ISession objects are cheap to create and dispose, because all of the model complexity is stored in the ISessionFactory and Configuration objects. As for Entity Framework, the ObjectContext/DbContext holds the configuration, model and acts as the Unit of Work, holding references to all of the known entity instances. This class is therefore not lightweight as its NHibernate counterpart and it is not uncommon to see examples where an instance is cached on a field. Mappings Both NHibernate and Entity Framework (Code First) support the use of POCOs to represent entities, no base classes are required (or even possible, in the case of NHibernate). As for mapping to and from the database, NHibernate supports three types of mappings: XML-based, which have the advantage of not tying the entity classes to a particular O/RM; the XML files can be deployed as files on the file system or as embedded resources in an assembly; Attribute-based, for keeping both the entities and database details on the same place at the expense of polluting the entity classes with NHibernate-specific attributes; Strongly-typed code-based, which allows dynamic creation of the model and strongly typing it, so that if, for example, a property name changes, the mapping will also be updated. Entity Framework can use: Attribute-based (although attributes cannot express all of the available possibilities – for example, cascading); Strongly-typed code mappings. Database Support With NHibernate you can use mostly any database you want, including: SQL Server; SQL Server Compact; SQL Server Azure; Oracle; DB2; PostgreSQL; MySQL; Sybase Adaptive Server/SQL Anywhere; Firebird; SQLLite; Informix; Any through OLE DB; Any through ODBC. Out of the box, Entity Framework only supports SQL Server, but a number of providers exist, both free and commercial, for some of the most used databases, such as Oracle and MySQL. See a list here. Inheritance Strategies Both NHibernate and Entity Framework support the three canonical inheritance strategies: Table Per Type Hierarchy (Single Table Inheritance), Table Per Type (Class Table Inheritance) and Table Per Concrete Type (Concrete Table Inheritance). Associations Regarding associations, both support one to one, one to many and many to many. However, NHibernate offers far more collection types: Bags of entities or values: unordered, possibly with duplicates; Lists of entities or values: ordered, indexed by a number column; Maps of entities or values: indexed by either an entity or any value; Sets of entities or values: unordered, no duplicates; Arrays of entities or values: indexed, immutable. Querying NHibernate exposes several querying APIs: LINQ is probably the most used nowadays, and really does not need to be introduced; Hibernate Query Language (HQL) is a database-agnostic, object-oriented SQL-alike language that exists since NHibernate’s creation and still offers the most advanced querying possibilities; well suited for dynamic queries, even if using string concatenation; Criteria API is an implementation of the Query Object pattern where you create a semi-abstract conceptual representation of the query you wish to execute by means of a class model; also a good choice for dynamic querying; Query Over offers a similar API to Criteria, but using strongly-typed LINQ expressions instead of strings; for this, although more refactor-friendlier that Criteria, it is also less suited for dynamic queries; SQL, including stored procedures, can also be used; Integration with Lucene.NET indexer is available. As for Entity Framework: LINQ to Entities is fully supported, and its implementation is considered very complete; it is the API of choice for most developers; Entity-SQL, HQL’s counterpart, is also an object-oriented, database-independent querying language that can be used for dynamic queries; SQL, of course, is also supported. Caching Both NHibernate and Entity Framework, of course, feature first-level cache. NHibernate also supports a second-level cache, that can be used among multiple ISessionFactorys, even in different processes/machines: Hashtable (in-memory); SysCache (uses ASP.NET as the cache provider); SysCache2 (same as above but with support for SQL Server SQL Dependencies); Prevalence; SharedCache; Memcached; Redis; NCache; Appfabric Caching. Out of the box, Entity Framework does not have any second-level cache mechanism, however, there are some public samples that show how we can add this. ID Generators NHibernate supports different ID generation strategies, coming from the database and otherwise: Identity (for SQL Server, MySQL, and databases who support identity columns); Sequence (for Oracle, PostgreSQL, and others who support sequences); Trigger-based; HiLo; Sequence HiLo (for databases that support sequences); Several GUID flavors, both in GUID as well as in string format; Increment (for single-user uses); Assigned (must know what you’re doing); Sequence-style (either uses an actual sequence or a single-column table); Table of ids; Pooled (similar to HiLo but stores high values in a table); Native (uses whatever mechanism the current database supports, identity or sequence). Entity Framework only supports: Identity generation; GUIDs; Assigned values. Properties NHibernate supports properties of entity types (one to one or many to one), collections (one to many or many to many) as well as scalars and enumerations. It offers a mechanism for having complex property types generated from the database, which even include support for querying. It also supports properties originated from SQL formulas. Entity Framework only supports scalars, entity types and collections. Enumerations support will come in the next version. Events and Interception NHibernate has a very rich event model, that exposes more than 20 events, either for synchronous pre-execution or asynchronous post-execution, including: Pre/Post-Load; Pre/Post-Delete; Pre/Post-Insert; Pre/Post-Update; Pre/Post-Flush. It also features interception of class instancing and SQL generation. As for Entity Framework, only two events exist: ObjectMaterialized (after loading an entity from the database); SavingChanges (before saving changes, which include deleting, inserting and updating). Tracking Changes For NHibernate as well as Entity Framework, all changes are tracked by their respective Unit of Work implementation. Entities can be attached and detached to it, Entity Framework does, however, also support self-tracking entities. Optimistic Concurrency Control NHibernate supports all of the imaginable scenarios: SQL Server’s ROWVERSION; Oracle’s ORA_ROWSCN; A column containing date and time; A column containing a version number; All/dirty columns comparison. Entity Framework is more focused on Entity Framework, so it only supports: SQL Server’s ROWVERSION; Comparing all/some columns. Batching NHibernate has full support for insertion batching, but only if the ID generator in use is not database-based (for example, it cannot be used with Identity), whereas Entity Framework has no batching at all. Cascading Both support cascading for collections and associations: when an entity is deleted, their conceptual children are also deleted. NHibernate also offers the possibility to set the foreign key column on children to NULL instead of removing them. Flushing Changes NHibernate’s ISession has a FlushMode property that can have the following values: Auto: changes are sent to the database when necessary, for example, if there are dirty instances of an entity type, and a query is performed against this entity type, or if the ISession is being disposed; Commit: changes are sent when committing the current transaction; Never: changes are only sent when explicitly calling Flush(). As for Entity Framework, changes have to be explicitly sent through a call to AcceptAllChanges()/SaveChanges(). Lazy Loading NHibernate supports lazy loading for Associated entities (one to one, many to one); Collections (one to many, many to many); Scalar properties (thing of BLOBs or CLOBs). Entity Framework only supports lazy loading for: Associated entities; Collections. Generating and Updating the Database Both NHibernate and Entity Framework Code First (with the Migrations API) allow creating the database model from the mapping and updating it if the mapping changes. Extensibility As you can guess, NHibernate is far more extensible than Entity Framework. Basically, everything can be extended, from ID generation, to LINQ to SQL transformation, HQL native SQL support, custom column types, custom association collections, SQL generation, supported databases, etc. With Entity Framework your options are more limited, at least, because practically no information exists as to what can be extended/changed. It features a provider model that can be extended to support any database. Integration With Other Microsoft APIs and Tools When it comes to integration with Microsoft technologies, it will come as no surprise that Entity Framework offers the best support. For example, the following technologies are fully supported: ASP.NET (through the EntityDataSource); ASP.NET Dynamic Data; WCF Data Services; WCF RIA Services; Visual Studio (through the integrated designer). Documentation This is another point where Entity Framework is superior: NHibernate lacks, for starters, an up to date API reference synchronized with its current version. It does have a community mailing list, blogs and wikis, although not much used. Entity Framework has a number of resources on MSDN and, of course, several forums and discussion groups exist. Conclusion Like I said, this is a personal list. I may come as a surprise to some that Entity Framework is so behind NHibernate in so many aspects, but it is true that NHibernate is much older and, due to its open-source nature, is not tied to product-specific timeframes and can thus evolve much more rapidly. I do like both, and I chose whichever is best for the job I have at hands. I am looking forward to the changes in EF5 which will add significant value to an already interesting product. So, what do you think? Did I forget anything important or is there anything else worth talking about? Looking forward for your comments!

    Read the article

  • Announcing Entity Framework Code-First (CTP5 release)

    - by ScottGu
    This week the data team released the CTP5 build of the new Entity Framework Code-First library.  EF Code-First enables a pretty sweet code-centric development workflow for working with data.  It enables you to: Develop without ever having to open a designer or define an XML mapping file Define model objects by simply writing “plain old classes” with no base classes required Use a “convention over configuration” approach that enables database persistence without explicitly configuring anything Optionally override the convention-based persistence and use a fluent code API to fully customize the persistence mapping I’m a big fan of the EF Code-First approach, and wrote several blog posts about it this summer: Code-First Development with Entity Framework 4 (July 16th) EF Code-First: Custom Database Schema Mapping (July 23rd) Using EF Code-First with an Existing Database (August 3rd) Today’s new CTP5 release delivers several nice improvements over the CTP4 build, and will be the last preview build of Code First before the final release of it.  We will ship the final EF Code First release in the first quarter of next year (Q1 of 2011).  It works with all .NET application types (including both ASP.NET Web Forms and ASP.NET MVC projects). Installing EF Code First You can install and use EF Code First CTP5 using one of two ways: Approach 1) By downloading and running a setup program.  Once installed you can reference the EntityFramework.dll assembly it provides within your projects.      or: Approach 2) By using the NuGet Package Manager within Visual Studio to download and install EF Code First within a project.  To do this, simply bring up the NuGet Package Manager Console within Visual Studio (View->Other Windows->Package Manager Console) and type “Install-Package EFCodeFirst”: Typing “Install-Package EFCodeFirst” within the Package Manager Console will cause NuGet to download the EF Code First package, and add it to your current project: Doing this will automatically add a reference to the EntityFramework.dll assembly to your project:   NuGet enables you to have EF Code First setup and ready to use within seconds.  When the final release of EF Code First ships you’ll also be able to just type “Update-Package EFCodeFirst” to update your existing projects to use the final release. EF Code First Assembly and Namespace The CTP5 release of EF Code First has an updated assembly name, and new .NET namespace: Assembly Name: EntityFramework.dll Namespace: System.Data.Entity These names match what we plan to use for the final release of the library. Nice New CTP5 Improvements The new CTP5 release of EF Code First contains a bunch of nice improvements and refinements. Some of the highlights include: Better support for Existing Databases Built-in Model-Level Validation and DataAnnotation Support Fluent API Improvements Pluggable Conventions Support New Change Tracking API Improved Concurrency Conflict Resolution Raw SQL Query/Command Support The rest of this blog post contains some more details about a few of the above changes. Better Support for Existing Databases EF Code First makes it really easy to create model layers that work against existing databases.  CTP5 includes some refinements that further streamline the developer workflow for this scenario. Below are the steps to use EF Code First to create a model layer for the Northwind sample database: Step 1: Create Model Classes and a DbContext class Below is all of the code necessary to implement a simple model layer using EF Code First that goes against the Northwind database: EF Code First enables you to use “POCO” – Plain Old CLR Objects – to represent entities within a database.  This means that you do not need to derive model classes from a base class, nor implement any interfaces or data persistence attributes on them.  This enables the model classes to be kept clean, easily testable, and “persistence ignorant”.  The Product and Category classes above are examples of POCO model classes. EF Code First enables you to easily connect your POCO model classes to a database by creating a “DbContext” class that exposes public properties that map to the tables within a database.  The Northwind class above illustrates how this can be done.  It is mapping our Product and Category classes to the “Products” and “Categories” tables within the database.  The properties within the Product and Category classes in turn map to the columns within the Products and Categories tables – and each instance of a Product/Category object maps to a row within the tables. The above code is all of the code required to create our model and data access layer!  Previous CTPs of EF Code First required an additional step to work against existing databases (a call to Database.Initializer<Northwind>(null) to tell EF Code First to not create the database) – this step is no longer required with the CTP5 release.  Step 2: Configure the Database Connection String We’ve written all of the code we need to write to define our model layer.  Our last step before we use it will be to setup a connection-string that connects it with our database.  To do this we’ll add a “Northwind” connection-string to our web.config file (or App.Config for client apps) like so:   <connectionStrings>          <add name="Northwind"          connectionString="data source=.\SQLEXPRESS;Integrated Security=SSPI;AttachDBFilename=|DataDirectory|\northwind.mdf;User Instance=true"          providerName="System.Data.SqlClient" />   </connectionStrings> EF “code first” uses a convention where DbContext classes by default look for a connection-string that has the same name as the context class.  Because our DbContext class is called “Northwind” it by default looks for a “Northwind” connection-string to use.  Above our Northwind connection-string is configured to use a local SQL Express database (stored within the \App_Data directory of our project).  You can alternatively point it at a remote SQL Server. Step 3: Using our Northwind Model Layer We can now easily query and update our database using the strongly-typed model layer we just built with EF Code First. The code example below demonstrates how to use LINQ to query for products within a specific product category.  This query returns back a sequence of strongly-typed Product objects that match the search criteria: The code example below demonstrates how we can retrieve a specific Product object, update two of its properties, and then save the changes back to the database: EF Code First handles all of the change-tracking and data persistence work for us, and allows us to focus on our application and business logic as opposed to having to worry about data access plumbing. Built-in Model Validation EF Code First allows you to use any validation approach you want when implementing business rules with your model layer.  This enables a great deal of flexibility and power. Starting with this week’s CTP5 release, EF Code First also now includes built-in support for both the DataAnnotation and IValidatorObject validation support built-into .NET 4.  This enables you to easily implement validation rules on your models, and have these rules automatically be enforced by EF Code First whenever you save your model layer.  It provides a very convenient “out of the box” way to enable validation within your applications. Applying DataAnnotations to our Northwind Model The code example below demonstrates how we could add some declarative validation rules to two of the properties of our “Product” model: We are using the [Required] and [Range] attributes above.  These validation attributes live within the System.ComponentModel.DataAnnotations namespace that is built-into .NET 4, and can be used independently of EF.  The error messages specified on them can either be explicitly defined (like above) – or retrieved from resource files (which makes localizing applications easy). Validation Enforcement on SaveChanges() EF Code-First (starting with CTP5) now automatically applies and enforces DataAnnotation rules when a model object is updated or saved.  You do not need to write any code to enforce this – this support is now enabled by default.  This new support means that the below code – which violates our above rules – will automatically throw an exception when we call the “SaveChanges()” method on our Northwind DbContext: The DbEntityValidationException that is raised when the SaveChanges() method is invoked contains a “EntityValidationErrors” property that you can use to retrieve the list of all validation errors that occurred when the model was trying to save.  This enables you to easily guide the user on how to fix them.  Note that EF Code-First will abort the entire transaction of changes if a validation rule is violated – ensuring that our database is always kept in a valid, consistent state. EF Code First’s validation enforcement works both for the built-in .NET DataAnnotation attributes (like Required, Range, RegularExpression, StringLength, etc), as well as for any custom validation rule you create by sub-classing the System.ComponentModel.DataAnnotations.ValidationAttribute base class. UI Validation Support A lot of our UI frameworks in .NET also provide support for DataAnnotation-based validation rules. For example, ASP.NET MVC, ASP.NET Dynamic Data, and Silverlight (via WCF RIA Services) all provide support for displaying client-side validation UI that honor the DataAnnotation rules applied to model objects. The screen-shot below demonstrates how using the default “Add-View” scaffold template within an ASP.NET MVC 3 application will cause appropriate validation error messages to be displayed if appropriate values are not provided: ASP.NET MVC 3 supports both client-side and server-side enforcement of these validation rules.  The error messages displayed are automatically picked up from the declarative validation attributes – eliminating the need for you to write any custom code to display them. Keeping things DRY The “DRY Principle” stands for “Do Not Repeat Yourself”, and is a best practice that recommends that you avoid duplicating logic/configuration/code in multiple places across your application, and instead specify it only once and have it apply everywhere. EF Code First CTP5 now enables you to apply declarative DataAnnotation validations on your model classes (and specify them only once) and then have the validation logic be enforced (and corresponding error messages displayed) across all applications scenarios – including within controllers, views, client-side scripts, and for any custom code that updates and manipulates model classes. This makes it much easier to build good applications with clean code, and to build applications that can rapidly iterate and evolve. Other EF Code First Improvements New to CTP5 EF Code First CTP5 includes a bunch of other improvements as well.  Below are a few short descriptions of some of them: Fluent API Improvements EF Code First allows you to override an “OnModelCreating()” method on the DbContext class to further refine/override the schema mapping rules used to map model classes to underlying database schema.  CTP5 includes some refinements to the ModelBuilder class that is passed to this method which can make defining mapping rules cleaner and more concise.  The ADO.NET Team blogged some samples of how to do this here. Pluggable Conventions Support EF Code First CTP5 provides new support that allows you to override the “default conventions” that EF Code First honors, and optionally replace them with your own set of conventions. New Change Tracking API EF Code First CTP5 exposes a new set of change tracking information that enables you to access Original, Current & Stored values, and State (e.g. Added, Unchanged, Modified, Deleted).  This support is useful in a variety of scenarios. Improved Concurrency Conflict Resolution EF Code First CTP5 provides better exception messages that allow access to the affected object instance and the ability to resolve conflicts using current, original and database values.  Raw SQL Query/Command Support EF Code First CTP5 now allows raw SQL queries and commands (including SPROCs) to be executed via the SqlQuery and SqlCommand methods exposed off of the DbContext.Database property.  The results of these method calls can be materialized into object instances that can be optionally change-tracked by the DbContext.  This is useful for a variety of advanced scenarios. Full Data Annotations Support EF Code First CTP5 now supports all standard DataAnnotations within .NET, and can use them both to perform validation as well as to automatically create the appropriate database schema when EF Code First is used in a database creation scenario.  Summary EF Code First provides an elegant and powerful way to work with data.  I really like it because it is extremely clean and supports best practices, while also enabling solutions to be implemented very, very rapidly.  The code-only approach of the library means that model layers end up being flexible and easy to customize. This week’s CTP5 release further refines EF Code First and helps ensure that it will be really sweet when it ships early next year.  I recommend using NuGet to install and give it a try today.  I think you’ll be pleasantly surprised by how awesome it is. Hope this helps, Scott

    Read the article

  • LLBLGen Pro feature highlights: automatic element name construction

    - by FransBouma
    (This post is part of a series of posts about features of the LLBLGen Pro system) One of the things one might take for granted but which has a huge impact on the time spent in an entity modeling environment is the way the system creates names for elements out of the information provided, in short: automatic element name construction. Element names are created in both directions of modeling: database first and model first and the more names the system can create for you without you having to rename them, the better. LLBLGen Pro has a rich, fine grained system for creating element names out of the meta-data available, which I'll describe more in detail below. First the model element related element naming features are highlighted, in the section Automatic model element naming features and after that I'll go more into detail about the relational model element naming features LLBLGen Pro has to offer in the section Automatic relational model element naming features. Automatic model element naming features When working database first, the element names in the model, e.g. entity names, entity field names and so on, are in general determined from the relational model element (e.g. table, table field) they're mapped on, as the model elements are reverse engineered from these relational model elements. It doesn't take rocket science to automatically name an entity Customer if the entity was created after reverse engineering a table named Customer. It gets a little trickier when the entity which was created by reverse engineering a table called TBL_ORDER_LINES has to be named 'OrderLine' automatically. Automatic model element naming also takes into effect with model first development, where some settings are used to provide you with a default name, e.g. in the case of navigator name creation when you create a new relationship. The features below are available to you in the Project Settings. Open Project Settings on a loaded project and navigate to Conventions -> Element Name Construction. Strippers! The above example 'TBL_ORDER_LINES' shows that some parts of the table name might not be needed for name creation, in this case the 'TBL_' prefix. Some 'brilliant' DBAs even add suffixes to table names, fragments you might not want to appear in the entity names. LLBLGen Pro offers you to define both prefix and suffix fragments to strip off of table, view, stored procedure, parameter, table field and view field names. In the example above, the fragment 'TBL_' is a good candidate for such a strip pattern. You can specify more than one pattern for e.g. the table prefix strip pattern, so even a really messy schema can still be used to produce clean names. Underscores Be Gone Another thing you might get rid of are underscores. After all, most naming schemes for entities and their classes use PasCal casing rules and don't allow for underscores to appear. LLBLGen Pro can automatically strip out underscores for you. It's an optional feature, so if you like the underscores, you're not forced to see them go: LLBLGen Pro will leave them alone when ordered to to so. PasCal everywhere... or not, your call LLBLGen Pro can automatically PasCal case names on word breaks. It determines word breaks in a couple of ways: a space marks a word break, an underscore marks a word break and a case difference marks a word break. It will remove spaces in all cases, and based on the underscore removal setting, keep or remove the underscores, and upper-case the first character of a word break fragment, and lower case the rest. Say, we keep the defaults, which is remove underscores and PasCal case always and strip the TBL_ fragment, we get with our example TBL_ORDER_LINES, after stripping TBL_ from the table name two word fragments: ORDER and LINES. The underscores are removed, the first character of each fragment is upper-cased, the rest lower-cased, so this results in OrderLines. Almost there! Pluralization and Singularization In general entity names are singular, like Customer or OrderLine so LLBLGen Pro offers a way to singularize the names. This will convert OrderLines, the result we got after the PasCal casing functionality, into OrderLine, exactly what we're after. Show me the patterns! There are other situations in which you want more flexibility. Say, you have an entity Customer and an entity Order and there's a foreign key constraint defined from the target of Order and the target of Customer. This foreign key constraint results in a 1:n relationship between the entities Customer and Order. A relationship has navigators mapped onto the relationship in both entities the relationship is between. For this particular relationship we'd like to have Customer as navigator in Order and Orders as navigator in Customer, so the relationship becomes Customer.Orders 1:n Order.Customer. To control the naming of these navigators for the various relationship types, LLBLGen Pro defines a set of patterns which allow you, using macros, to define how the auto-created navigator names will look like. For example, if you rather have Customer.OrderCollection, you can do so, by changing the pattern from {$EndEntityName$P} to {$EndEntityName}Collection. The $P directive makes sure the name is pluralized, which is not what you want if you're going for <EntityName>Collection, hence it's removed. When working model first, it's a given you'll create foreign key fields along the way when you define relationships. For example, you've defined two entities: Customer and Order, and they have their fields setup properly. Now you want to define a relationship between them. This will automatically create a foreign key field in the Order entity, which reflects the value of the PK field in Customer. (No worries if you hate the foreign key fields in your classes, on NHibernate and EF these can be hidden in the generated code if you want to). A specific pattern is available for you to direct LLBLGen Pro how to name this foreign key field. For example, if all your entities have Id as PK field, you might want to have a different name than Id as foreign key field. In our Customer - Order example, you might want to have CustomerId instead as foreign key name in Order. The pattern for foreign key fields gives you that freedom. Abbreviations... make sense of OrdNr and friends I already described word breaks in the PasCal casing paragraph, how they're used for the PasCal casing in the constructed name. Word breaks are used for another neat feature LLBLGen Pro has to offer: abbreviation support. Burt, your friendly DBA in the dungeons below the office has a hate-hate relationship with his keyboard: he can't stand it: typing is something he avoids like the plague. This has resulted in tables and fields which have names which are very short, but also very unreadable. Example: our TBL_ORDER_LINES example has a lovely field called ORD_NR. What you would like to see in your fancy new OrderLine entity mapped onto this table is a field called OrderNumber, not a field called OrdNr. What you also like is to not have to rename that field manually. There are better things to do with your time, after all. LLBLGen Pro has you covered. All it takes is to define some abbreviation - full word pairs and during reverse engineering model elements from tables/views, LLBLGen Pro will take care of the rest. For the ORD_NR field, you need two values: ORD as abbreviation and Order as full word, and NR as abbreviation and Number as full word. LLBLGen Pro will now convert every word fragment found with the word breaks which matches an abbreviation to the given full word. They're case sensitive and can be found in the Project Settings: Navigate to Conventions -> Element Name Construction -> Abbreviations. Automatic relational model element naming features Not everyone works database first: it may very well be the case you start from scratch, or have to add additional tables to an existing database. For these situations, it's key you have the flexibility that you can control the created table names and table fields without any work: let the designer create these names based on the entity model you defined and a set of rules. LLBLGen Pro offers several features in this area, which are described in more detail below. These features are found in Project Settings: navigate to Conventions -> Model First Development. Underscores, welcome back! Not every database is case insensitive, and not every organization requires PasCal cased table/field names, some demand all lower or all uppercase names with underscores at word breaks. Say you create an entity model with an entity called OrderLine. You work with Oracle and your organization requires underscores at word breaks: a table created from OrderLine should be called ORDER_LINE. LLBLGen Pro allows you to do that: with a simple checkbox you can order LLBLGen Pro to insert an underscore at each word break for the type of database you're working with: case sensitive or case insensitive. Checking the checkbox Insert underscore at word break case insensitive dbs will let LLBLGen Pro create a table from the entity called Order_Line. Half-way there, as there are still lower case characters there and you need all caps. No worries, see below Casing directives so everyone can sleep well at night For case sensitive databases and case insensitive databases there is one setting for each of them which controls the casing of the name created from a model element (e.g. a table created from an entity definition using the auto-mapping feature). The settings can have the following values: AsProjectElement, AllUpperCase or AllLowerCase. AsProjectElement is the default, and it keeps the casing as-is. In our example, we need to get all upper case characters, so we select AllUpperCase for the setting for case sensitive databases. This will produce the name ORDER_LINE. Sequence naming after a pattern Some databases support sequences, and using model-first development it's key to have sequences, when needed, to be created automatically and if possible using a name which shows where they're used. Say you have an entity Order and you want to have the PK values be created by the database using a sequence. The database you're using supports sequences (e.g. Oracle) and as you want all numeric PK fields to be sequenced, you have enabled this by the setting Auto assign sequences to integer pks. When you're using LLBLGen Pro's auto-map feature, to create new tables and constraints from the model, it will create a new table, ORDER, based on your settings I previously discussed above, with a PK field ID and it also creates a sequence, SEQ_ORDER, which is auto-assigns to the ID field mapping. The name of the sequence is created by using a pattern, defined in the Model First Development setting Sequence pattern, which uses plain text and macros like with the other patterns previously discussed. Grouping and schemas When you start from scratch, and you're working model first, the tables created by LLBLGen Pro will be in a catalog and / or schema created by LLBLGen Pro as well. If you use LLBLGen Pro's grouping feature, which allows you to group entities and other model elements into groups in the project (described in a future blog post), you might want to have that group name reflected in the schema name the targets of the model elements are in. Say you have a model with a group CRM and a group HRM, both with entities unique for these groups, e.g. Employee in HRM, Customer in CRM. When auto-mapping this model to create tables, you might want to have the table created for Employee in the HRM schema but the table created for Customer in the CRM schema. LLBLGen Pro will do just that when you check the setting Set schema name after group name to true (default). This gives you total control over where what is placed in the database from your model. But I want plural table names... and TBL_ prefixes! For now we follow best practices which suggest singular table names and no prefixes/suffixes for names. Of course that won't keep everyone happy, so we're looking into making it possible to have that in a future version. Conclusion LLBLGen Pro offers a variety of options to let the modeling system do as much work for you as possible. Hopefully you enjoyed this little highlight post and that it has given you new insights in the smaller features available to you in LLBLGen Pro, ones you might not have thought off in the first place. Enjoy!

    Read the article

  • Oracle BI Server Modeling, Part 1- Designing a Query Factory

    - by bob.ertl(at)oracle.com
      Welcome to Oracle BI Development's BI Foundation blog, focused on helping you get the most value from your Oracle Business Intelligence Enterprise Edition (BI EE) platform deployments.  In my first series of posts, I plan to show developers the concepts and best practices for modeling in the Common Enterprise Information Model (CEIM), the semantic layer of Oracle BI EE.  In this segment, I will lay the groundwork for the modeling concepts.  First, I will cover the big picture of how the BI Server fits into the system, and how the CEIM controls the query processing. Oracle BI EE Query Cycle The purpose of the Oracle BI Server is to bridge the gap between the presentation services and the data sources.  There are typically a variety of data sources in a variety of technologies: relational, normalized transaction systems; relational star-schema data warehouses and marts; multidimensional analytic cubes and financial applications; flat files, Excel files, XML files, and so on. Business datasets can reside in a single type of source, or, most of the time, are spread across various types of sources. Presentation services users are generally business people who need to be able to query that set of sources without any knowledge of technologies, schemas, or how sources are organized in their company. They think of business analysis in terms of measures with specific calculations, hierarchical dimensions for breaking those measures down, and detailed reports of the business transactions themselves.  Most of them create queries without knowing it, by picking a dashboard page and some filters.  Others create their own analysis by selecting metrics and dimensional attributes, and possibly creating additional calculations. The BI Server bridges that gap from simple business terms to technical physical queries by exposing just the business focused measures and dimensional attributes that business people can use in their analyses and dashboards.   After they make their selections and start the analysis, the BI Server plans the best way to query the data sources, writes the optimized sequence of physical queries to those sources, post-processes the results, and presents them to the client as a single result set suitable for tables, pivots and charts. The CEIM is a model that controls the processing of the BI Server.  It provides the subject areas that presentation services exposes for business users to select simplified metrics and dimensional attributes for their analysis.  It models the mappings to the physical data access, the calculations and logical transformations, and the data access security rules.  The CEIM consists of metadata stored in the repository, authored by developers using the Administration Tool client.     Presentation services and other query clients create their queries in BI EE's SQL-92 language, called Logical SQL or LSQL.  The API simply uses ODBC or JDBC to pass the query to the BI Server.  Presentation services writes the LSQL query in terms of the simplified objects presented to the users.  The BI Server creates a query plan, and rewrites the LSQL into fully-detailed SQL or other languages suitable for querying the physical sources.  For example, the LSQL on the left below was rewritten into the physical SQL for an Oracle 11g database on the right. Logical SQL   Physical SQL SELECT "D0 Time"."T02 Per Name Month" saw_0, "D4 Product"."P01  Product" saw_1, "F2 Units"."2-01  Billed Qty  (Sum All)" saw_2 FROM "Sample Sales" ORDER BY saw_0, saw_1       WITH SAWITH0 AS ( select T986.Per_Name_Month as c1, T879.Prod_Dsc as c2,      sum(T835.Units) as c3, T879.Prod_Key as c4 from      Product T879 /* A05 Product */ ,      Time_Mth T986 /* A08 Time Mth */ ,      FactsRev T835 /* A11 Revenue (Billed Time Join) */ where ( T835.Prod_Key = T879.Prod_Key and T835.Bill_Mth = T986.Row_Wid) group by T879.Prod_Dsc, T879.Prod_Key, T986.Per_Name_Month ) select SAWITH0.c1 as c1, SAWITH0.c2 as c2, SAWITH0.c3 as c3 from SAWITH0 order by c1, c2   Probably everybody reading this blog can write SQL or MDX.  However, the trick in designing the CEIM is that you are modeling a query-generation factory.  Rather than hand-crafting individual queries, you model behavior and relationships, thus configuring the BI Server machinery to manufacture millions of different queries in response to random user requests.  This mass production requires a different mindset and approach than when you are designing individual SQL statements in tools such as Oracle SQL Developer, Oracle Hyperion Interactive Reporting (formerly Brio), or Oracle BI Publisher.   The Structure of the Common Enterprise Information Model (CEIM) The CEIM has a unique structure specifically for modeling the relationships and behaviors that fill the gap from logical user requests to physical data source queries and back to the result.  The model divides the functionality into three specialized layers, called Presentation, Business Model and Mapping, and Physical, as shown below. Presentation services clients can generally only see the presentation layer, and the objects in the presentation layer are normally the only ones used in the LSQL request.  When a request comes into the BI Server from presentation services or another client, the relationships and objects in the model allow the BI Server to select the appropriate data sources, create a query plan, and generate the physical queries.  That's the left to right flow in the diagram below.  When the results come back from the data source queries, the right to left relationships in the model show how to transform the results and perform any final calculations and functions that could not be pushed down to the databases.   Business Model Think of the business model as the heart of the CEIM you are designing.  This is where you define the analytic behavior seen by the users, and the superset library of metric and dimension objects available to the user community as a whole.  It also provides the baseline business-friendly names and user-readable dictionary.  For these reasons, it is often called the "logical" model--it is a virtual database schema that persists no data, but can be queried as if it is a database. The business model always has a dimensional shape (more on this in future posts), and its simple shape and terminology hides the complexity of the source data models. Besides hiding complexity and normalizing terminology, this layer adds most of the analytic value, as well.  This is where you define the rich, dimensional behavior of the metrics and complex business calculations, as well as the conformed dimensions and hierarchies.  It contributes to the ease of use for business users, since the dimensional metric definitions apply in any context of filters and drill-downs, and the conformed dimensions enable dashboard-wide filters and guided analysis links that bring context along from one page to the next.  The conformed dimensions also provide a key to hiding the complexity of many sources, including federation of different databases, behind the simple business model. Note that the expression language in this layer is LSQL, so that any expression can be rewritten into any data source's query language at run time.  This is important for federation, where a given logical object can map to several different physical objects in different databases.  It is also important to portability of the CEIM to different database brands, which is a key requirement for Oracle's BI Applications products. Your requirements process with your user community will mostly affect the business model.  This is where you will define most of the things they specifically ask for, such as metric definitions.  For this reason, many of the best-practice methodologies of our consulting partners start with the high-level definition of this layer. Physical Model The physical model connects the business model that meets your users' requirements to the reality of the data sources you have available. In the query factory analogy, think of the physical layer as the bill of materials for generating physical queries.  Every schema, table, column, join, cube, hierarchy, etc., that will appear in any physical query manufactured at run time must be modeled here at design time. Each physical data source will have its own physical model, or "database" object in the CEIM.  The shape of each physical model matches the shape of its physical source.  In other words, if the source is normalized relational, the physical model will mimic that normalized shape.  If it is a hypercube, the physical model will have a hypercube shape.  If it is a flat file, it will have a denormalized tabular shape. To aid in query optimization, the physical layer also tracks the specifics of the database brand and release.  This allows the BI Server to make the most of each physical source's distinct capabilities, writing queries in its syntax, and using its specific functions. This allows the BI Server to push processing work as deep as possible into the physical source, which minimizes data movement and takes full advantage of the database's own optimizer.  For most data sources, native APIs are used to further optimize performance and functionality. The value of having a distinct separation between the logical (business) and physical models is encapsulation of the physical characteristics.  This encapsulation is another enabler of packaged BI applications and federation.  It is also key to hiding the complex shapes and relationships in the physical sources from the end users.  Consider a routine drill-down in the business model: physically, it can require a drill-through where the first query is MDX to a multidimensional cube, followed by the drill-down query in SQL to a normalized relational database.  The only difference from the user's point of view is that the 2nd query added a more detailed dimension level column - everything else was the same. Mappings Within the Business Model and Mapping Layer, the mappings provide the binding from each logical column and join in the dimensional business model, to each of the objects that can provide its data in the physical layer.  When there is more than one option for a physical source, rules in the mappings are applied to the query context to determine which of the data sources should be hit, and how to combine their results if more than one is used.  These rules specify aggregate navigation, vertical partitioning (fragmentation), and horizontal partitioning, any of which can be federated across multiple, heterogeneous sources.  These mappings are usually the most sophisticated part of the CEIM. Presentation You might think of the presentation layer as a set of very simple relational-like views into the business model.  Over ODBC/JDBC, they present a relational catalog consisting of databases, tables and columns.  For business users, presentation services interprets these as subject areas, folders and columns, respectively.  (Note that in 10g, subject areas were called presentation catalogs in the CEIM.  In this blog, I will stick to 11g terminology.)  Generally speaking, presentation services and other clients can query only these objects (there are exceptions for certain clients such as BI Publisher and Essbase Studio). The purpose of the presentation layer is to specialize the business model for different categories of users.  Based on a user's role, they will be restricted to specific subject areas, tables and columns for security.  The breakdown of the model into multiple subject areas organizes the content for users, and subjects superfluous to a particular business role can be hidden from that set of users.  Customized names and descriptions can be used to override the business model names for a specific audience.  Variables in the object names can be used for localization. For these reasons, you are better off thinking of the tables in the presentation layer as folders than as strict relational tables.  The real semantics of tables and how they function is in the business model, and any grouping of columns can be included in any table in the presentation layer.  In 11g, an LSQL query can also span multiple presentation subject areas, as long as they map to the same business model. Other Model Objects There are some objects that apply to multiple layers.  These include security-related objects, such as application roles, users, data filters, and query limits (governors).  There are also variables you can use in parameters and expressions, and initialization blocks for loading their initial values on a static or user session basis.  Finally, there are Multi-User Development (MUD) projects for developers to check out units of work, and objects for the marketing feature used by our packaged customer relationship management (CRM) software.   The Query Factory At this point, you should have a grasp on the query factory concept.  When developing the CEIM model, you are configuring the BI Server to automatically manufacture millions of queries in response to random user requests. You do this by defining the analytic behavior in the business model, mapping that to the physical data sources, and exposing it through the presentation layer's role-based subject areas. While configuring mass production requires a different mindset than when you hand-craft individual SQL or MDX statements, it builds on the modeling and query concepts you already understand. The following posts in this series will walk through the CEIM modeling concepts and best practices in detail.  We will initially review dimensional concepts so you can understand the business model, and then present a pattern-based approach to learning the mappings from a variety of physical schema shapes and deployments to the dimensional model.  Along the way, we will also present the dimensional calculation template, and learn how to configure the many additivity patterns.

    Read the article

  • Sync database with filter using SyncOrchestrator with Sync Framework 2.0

    - by Flo
    Hi, I want to synchronize two SQL databases. But since one of the databases only requires a subset of the data I am looking for a filter option. Is there a possibility to add a Filter to the SyncOrchestrator or do I have to add the filter to the SyncProvider? According to this: http://social.microsoft.com/Forums/en-US/uklaunch2007ado.net/thread/35d4deb8-a861-4fe3-a395-d175e14c353f it is not possible to filter with the DbSyncProvider. Quote: "I understand your scenario, and the hebavior of the DbSyncProvider is due to the current limitation. DbSyncProvider is built on top of the Microsoft Sync Framework that can support filtering. Unfortunately, DbSyncProvider does not yet." But that post is quite old, maybe that has changed now. I am working with this example at the moment: http://msdn.microsoft.com/en-us/library/cc807255.aspx but I can't figure out how to add filtering here.

    Read the article

< Previous Page | 37 38 39 40 41 42 43 44 45 46 47 48  | Next Page >