Search Results

Search found 772 results on 31 pages for 'ordering'.

Page 8/31 | < Previous Page | 4 5 6 7 8 9 10 11 12 13 14 15  | Next Page >

  • Why is UDP + a software reliable ordering system faster than TCP?

    - by Ricket
    Some games today use a network system that transmits messages over UDP, and ensures that the messages are reliable and ordered. For example, RakNet is a popular game network engine. It uses only UDP for its connections, and has a whole system to ensure that packets can be reliable and ordered if you so choose. My basic question is, what's up with that? Isn't TCP the same thing as ordered, reliable UDP? What makes it so much slower that people have to basically reinvent the wheel?

    Read the article

  • Create ordering in a MySQL table without using a number (because then it's hard to put something in

    - by user347256
    I have a long list of items (say, a few million items) in a mysql table, let's call it mytable and it has the field mytable.itemid. The items are given an order, and can be re=ordered by the user by drag and drop. If I add a field called mytable.order and just put numbers in them, it creates problems: what if I want to move an item between 2 other items? Then all the order fields have to be updated? That seems like a nightmare. Is there a (scalable) way to add order to a table that is different from just giving every item a number, order by that, and do loads of SQL queries everytime the order is changed?

    Read the article

  • How does one find out which application is associated with an indicator icon in Ubuntu 12.04?

    - by Amos Annoy
    It is trivial to do this in Ubuntu 10.04. The question is specific to Ubuntu 12.04. This has nothing to do (does it?) with right click. How can an indicator's icon in Ubuntu 12.04 be matched with the program responsible for it's manifestation on the top panel? A list of running applications can include all processes using System Monitor. How is the correct matching process found for an indicator? (the examination of SM points out a rather poignant factor in the faster depletion and shortened run time on battery - the ambient quiescent CPU rate in 12.04 is now well over 20% when previously it was well under 10% in 10.04, between 5% and 7%!) (I have a problem with the battery indicator - it sometimes has % and other times hh:mm - it is necessary to know the ap. & v. to get more info on controlling same. ditto: There are issues with other indicator aps.) Details from: How can I find Application Indicator ID's? suggests looking at: file:///usr/share/indicator-application/ordering-override.keyfile [Ordering Index Overrides] nm-applet=1 gnome-power-manager=2 ibus=3 gst-keyboard-xkb=4 gsd-keyboard-xkb=5 which solves the battery identification, and presumably nm is NetworkManager for the rf icon, but the envelope, blue tooth and speaker indicator aps. are still a mystery. (Also, the ordering is not correlated.) Mind you, it was simple in the past to simply right click to get the About option to find the ap. & v. info. browsing around and about: file:///usr/share/indicator-application/ordering-override.keyfile examined: file:///usr/share/indicators file:///usr/share/indicators/messages/applications/ ... perhaps?/presumably? the information sought may be buried in file:///usr/share/indicators

    Read the article

  • Access to the Technology Software

    - by rituchhibber
    The Technology Program software is available to Oracle partners, free of charge, for demonstration and development purposes according to the terms of the OPN Agreement. Instructions for Ordering Software: Download software via the Oracle Software Delivery Cloud website. Downloads are available in most countries. To determine if downloads are available within your country click on the Download software link. Request a physical shipment of the software by downloading and completing the Development and Demonstration Ordering Document (XLS). Should you have additional questions or need assistance in completing the ordering document, please contact your local Partner Business Center. Incomplete orders will not be processed.

    Read the article

  • Parallelism in .NET – Part 7, Some Differences between PLINQ and LINQ to Objects

    - by Reed
    In my previous post on Declarative Data Parallelism, I mentioned that PLINQ extends LINQ to Objects to support parallel operations.  Although nearly all of the same operations are supported, there are some differences between PLINQ and LINQ to Objects.  By introducing Parallelism to our declarative model, we add some extra complexity.  This, in turn, adds some extra requirements that must be addressed. In order to illustrate the main differences, and why they exist, let’s begin by discussing some differences in how the two technologies operate, and look at the underlying types involved in LINQ to Objects and PLINQ . LINQ to Objects is mainly built upon a single class: Enumerable.  The Enumerable class is a static class that defines a large set of extension methods, nearly all of which work upon an IEnumerable<T>.  Many of these methods return a new IEnumerable<T>, allowing the methods to be chained together into a fluent style interface.  This is what allows us to write statements that chain together, and lead to the nice declarative programming model of LINQ: double min = collection .Where(item => item.SomeProperty > 6 && item.SomeProperty < 24) .Min(item => item.PerformComputation()); .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } Other LINQ variants work in a similar fashion.  For example, most data-oriented LINQ providers are built upon an implementation of IQueryable<T>, which allows the database provider to turn a LINQ statement into an underlying SQL query, to be performed directly on the remote database. PLINQ is similar, but instead of being built upon the Enumerable class, most of PLINQ is built upon a new static class: ParallelEnumerable.  When using PLINQ, you typically begin with any collection which implements IEnumerable<T>, and convert it to a new type using an extension method defined on ParallelEnumerable: AsParallel().  This method takes any IEnumerable<T>, and converts it into a ParallelQuery<T>, the core class for PLINQ.  There is a similar ParallelQuery class for working with non-generic IEnumerable implementations. This brings us to our first subtle, but important difference between PLINQ and LINQ – PLINQ always works upon specific types, which must be explicitly created. Typically, the type you’ll use with PLINQ is ParallelQuery<T>, but it can sometimes be a ParallelQuery or an OrderedParallelQuery<T>.  Instead of dealing with an interface, implemented by an unknown class, we’re dealing with a specific class type.  This works seamlessly from a usage standpoint – ParallelQuery<T> implements IEnumerable<T>, so you can always “switch back” to an IEnumerable<T>.  The difference only arises at the beginning of our parallelization.  When we’re using LINQ, and we want to process a normal collection via PLINQ, we need to explicitly convert the collection into a ParallelQuery<T> by calling AsParallel().  There is an important consideration here – AsParallel() does not need to be called on your specific collection, but rather any IEnumerable<T>.  This allows you to place it anywhere in the chain of methods involved in a LINQ statement, not just at the beginning.  This can be useful if you have an operation which will not parallelize well or is not thread safe.  For example, the following is perfectly valid, and similar to our previous examples: double min = collection .AsParallel() .Select(item => item.SomeOperation()) .Where(item => item.SomeProperty > 6 && item.SomeProperty < 24) .Min(item => item.PerformComputation()); However, if SomeOperation() is not thread safe, we could just as easily do: double min = collection .Select(item => item.SomeOperation()) .AsParallel() .Where(item => item.SomeProperty > 6 && item.SomeProperty < 24) .Min(item => item.PerformComputation()); In this case, we’re using standard LINQ to Objects for the Select(…) method, then converting the results of that map routine to a ParallelQuery<T>, and processing our filter (the Where method) and our aggregation (the Min method) in parallel. PLINQ also provides us with a way to convert a ParallelQuery<T> back into a standard IEnumerable<T>, forcing sequential processing via standard LINQ to Objects.  If SomeOperation() was thread-safe, but PerformComputation() was not thread-safe, we would need to handle this by using the AsEnumerable() method: double min = collection .AsParallel() .Select(item => item.SomeOperation()) .Where(item => item.SomeProperty > 6 && item.SomeProperty < 24) .AsEnumerable() .Min(item => item.PerformComputation()); Here, we’re converting our collection into a ParallelQuery<T>, doing our map operation (the Select(…) method) and our filtering in parallel, then converting the collection back into a standard IEnumerable<T>, which causes our aggregation via Min() to be performed sequentially. This could also be written as two statements, as well, which would allow us to use the language integrated syntax for the first portion: var tempCollection = from item in collection.AsParallel() let e = item.SomeOperation() where (e.SomeProperty > 6 && e.SomeProperty < 24) select e; double min = tempCollection.AsEnumerable().Min(item => item.PerformComputation()); This allows us to use the standard LINQ style language integrated query syntax, but control whether it’s performed in parallel or serial by adding AsParallel() and AsEnumerable() appropriately. The second important difference between PLINQ and LINQ deals with order preservation.  PLINQ, by default, does not preserve the order of of source collection. This is by design.  In order to process a collection in parallel, the system needs to naturally deal with multiple elements at the same time.  Maintaining the original ordering of the sequence adds overhead, which is, in many cases, unnecessary.  Therefore, by default, the system is allowed to completely change the order of your sequence during processing.  If you are doing a standard query operation, this is usually not an issue.  However, there are times when keeping a specific ordering in place is important.  If this is required, you can explicitly request the ordering be preserved throughout all operations done on a ParallelQuery<T> by using the AsOrdered() extension method.  This will cause our sequence ordering to be preserved. For example, suppose we wanted to take a collection, perform an expensive operation which converts it to a new type, and display the first 100 elements.  In LINQ to Objects, our code might look something like: // Using IEnumerable<SourceClass> collection IEnumerable<ResultClass> results = collection .Select(e => e.CreateResult()) .Take(100); If we just converted this to a parallel query naively, like so: IEnumerable<ResultClass> results = collection .AsParallel() .Select(e => e.CreateResult()) .Take(100); We could very easily get a very different, and non-reproducable, set of results, since the ordering of elements in the input collection is not preserved.  To get the same results as our original query, we need to use: IEnumerable<ResultClass> results = collection .AsParallel() .AsOrdered() .Select(e => e.CreateResult()) .Take(100); This requests that PLINQ process our sequence in a way that verifies that our resulting collection is ordered as if it were processed serially.  This will cause our query to run slower, since there is overhead involved in maintaining the ordering.  However, in this case, it is required, since the ordering is required for correctness. PLINQ is incredibly useful.  It allows us to easily take nearly any LINQ to Objects query and run it in parallel, using the same methods and syntax we’ve used previously.  There are some important differences in operation that must be considered, however – it is not a free pass to parallelize everything.  When using PLINQ in order to parallelize your routines declaratively, the same guideline I mentioned before still applies: Parallelization is something that should be handled with care and forethought, added by design, and not just introduced casually.

    Read the article

  • Hibernate: order multiple one-to-many relations

    - by Markos Fragkakis
    I have a search screen, using JSF, JBoss Seam and Hibernate underneath. There are columns for A, B and C, where the relations are as follows: A (1< -- ) B (1< -- ) C A has a List< B and B has a List< C (both relations are one-to-many). The UI table supports ordering by any column (ASC or DESC), so I want the results of the query to be ordered. This is the reason I used Lists in the model. However, I got an exception that Hibernate cannot eagerly fetch multiple bags (it considers both lists to be bags). There is an interesting blog post here, and they identify the following solutions: Use @IndexColumn annotation (there is none in my DB, and what's more, I want the position of results to be determined by the ordering, not by an index column) Fetch lazily (for performance reasons, I need eager fetching) Change List to Set So, I changed the List to Set, which by the way is more correct, model-wise. First, if don't use @OrderBy, the PersistentSet returned by Hibernate wraps a HashSet, which has no ordering. Second, If I do use @OrderBy, the PersistentSet wraps a LinkedHashSet, which is what I would like, but the OrderBy property is hardcoded, so all other ordering I perform through the UI comes after it. I tried again with Sets, and used SortedSet (and its implementation, TreeSet), but I have some issues: I want ordering to take place in the DB, and not in-memory, which is what TreeSet does (either through a Comparator, or through the Comparable interface of the elements). Second, I found that there is the Hibernate annotation @Sort, which has a SortOrder.UNSORTED and you can also set a Comparator. I still haven't managed to make it compile, but I am still not convinced it is what I need. Any advice?

    Read the article

  • Popularity Algorithm - SQL / Django

    - by RadiantHex
    Hi folks, I've been looking into popularity algorithms used on sites such as Reddit, Digg and even Stackoverflow. Reddit algorithm: t = (time of entry post) - (Dec 8, 2005) x = upvotes - downvotes y = {1 if x > 0, 0 if x = 0, -1 if x < 0) z = {1 if x < 0, otherwise x} log(z) + (y * t)/45000 I have always performed simple ordering within SQL, I'm wondering how I should deal with such ordering. Should it be used to define a table, or could I build an SQL with the ordering within the formula (without hindering performance)? I am also wondering, if it is possible to use multiple ordering algorithms in different occasions, without incurring into performance problems. I'm using Django and PostgreSQL. Help would be much appreciated! ^^

    Read the article

  • heterogeneous comparisons in python3

    - by Matt Anderson
    I'm 99+% still using python 2.x, but I'm trying to think ahead to the day when I switch. So, I know that using comparison operators (less/greater than, or equal to) on heterogeneous types that don't have a natural ordering is no longer supported in python3.x -- instead of some consistent (but arbitrary) result we raise TypeError instead. I see the logic in that, and even mostly think its a good thing. Consistency and refusing to guess is a virtue. But what if you essentially want the python2.x behavior? What's the best way to go about getting it? For fun (more or less) I was recently implementing a Skip List, a data structure that keeps its elements sorted. I wanted to use heterogeneous types as keys in the data structure, and I've got to compare keys to one another as I walk the data structure. The python2.x way of comparing makes this really convenient -- you get an understandable ordering amongst elements that have a natural ordering, and some ordering amongst those that don't. Consistently using a sort/comparison key like (type(obj).__name__, obj) has the disadvantage of not interleaving the objects that do have a natural ordering; you get all your floats clustered together before your ints, and your str-derived class separates from your strs. I came up with the following: import operator def hetero_sort_key(obj): cls = type(obj) return (cls.__name__+'_'+cls.__module__, obj) def make_hetero_comparitor(fn): def comparator(a, b): try: return fn(a, b) except TypeError: return fn(hetero_sort_key(a), hetero_sort_key(b)) return comparator hetero_lt = make_hetero_comparitor(operator.lt) hetero_gt = make_hetero_comparitor(operator.gt) hetero_le = make_hetero_comparitor(operator.le) hetero_ge = make_hetero_comparitor(operator.gt) Is there a better way? I suspect one could construct a corner case that this would screw up -- a situation where you can compare type A to B and type A to C, but where B and C raise TypeError when compared, and you can end up with something illogical like a > b, a < c, and yet b > c (because of how their class names sorted). I don't know how likely it is that you'd run into this in practice.

    Read the article

  • Simple PHP query question: LIKE

    - by pg
    When I replace $ordering = "apples, bananas, cranberries, grapes"; with $ordering = "apples, bananas, grapes"; I no longer want cranberries to be returned by my query, which I've written out like this: $query = "SELECT * from dbname where FruitName LIKE '$ordering'"; Of Course this doesn't work, because I used LIKE wrong. I've read through various manuals that describe how to use LIKE and it doesn't quite make sense to me. If I change the end of the db to "LIKE "apples"" that works for limiting it to just apples. Do I have to explode the ordering on the ", " or is there a way to do this in the query?

    Read the article

  • Scala importing a file in all files of a package

    - by Core_Dumped
    I need to use an implicit ordering that has been defined in an object in a file abc in the following way: object abc{ implicit def localTimeOrdering: Ordering[LocalDate] = Ordering.fromLessThan(_.isBefore(_)) } So, I make a package object xyz inside a file 'package.scala' that in turn is in the package 'xyz' that has files in which I need the implicit ordering to be applicable. I write something like this: package object xyz{ import abc._ } It does not seem to work. If I manually write the implicit definition statement inside the package object, it works perfectly. What is the correct way to import the object (abc) such that all of its objects/classes/definitions can be used in my entire package 'xyz' ?

    Read the article

  • Oracle’s Visual CRM Solution

    Visual CRM adds the powerful visualization and document centric collaboration capabilities of Oracle’s AutoVue to Oracle’s best-in-class CRM solutions. By introducing a visual aspect to call center, field service, and ordering processes, Visual CRM helps teams provide faster responses to customer issues, optimize field service performance, and shorten ordering cycles while minimizing order errors.With Visual CRM, organizations can achieve improved customer service levels and field service operations which help drive margin, top line revenue, and customer retention.

    Read the article

  • Speeding up ROW_NUMBER in SQL Server

    - by BlueRaja
    We have a number of machines which record data into a database at sporadic intervals. For each record, I'd like to obtain the time period between this recording and the previous recording. I can do this using ROW_NUMBER as follows: WITH TempTable AS ( SELECT *, ROW_NUMBER() OVER (PARTITION BY Machine_ID ORDER BY Date_Time) AS Ordering FROM dbo.DataTable ) SELECT [Current].*, Previous.Date_Time AS PreviousDateTime FROM TempTable AS [Current] INNER JOIN TempTable AS Previous ON [Current].Machine_ID = Previous.Machine_ID AND Previous.Ordering = [Current].Ordering + 1 The problem is, it goes really slow (several minutes on a table with about 10k entries) - I tried creating separate indicies on Machine_ID and Date_Time, and a single joined-index, but nothing helps. Is there anyway to rewrite this query to go faster?

    Read the article

  • When are SQL views appropriate in ASP.net MVC?

    - by sslepian
    I've got a table called Protocol, a table called Eligibility, and a Protocol_Eligibilty table that maps the two together (a many to many relationship). If I wanted to make a perfect copy of an entry in the Protocol table, and create all the needed mappings in the Protocol_Eligibility table, would using an SQL view be helpful, from a performance standpoint? Protocol will have around 1000 rows, Eligibility will have about 200, and I expect each Protocol to map to about 10 Eligibility rows and each Eligibility to map to over 100 rows in Protocol. Here's how I'm doing this with the view: var pel_original = (from pel in _documentDataModel.Protocol_Eligibility_View where pel.pid == id select pel); Protocol_Eligibility newEligibility; foreach (var pel_item in pel_original) { newEligibility = new Protocol_Eligibility(); newEligibility.Eligibility = (from pel in _documentDataModel.Eligibility where pel.ID == pel_item.eid select pel).First(); newEligibility.Protocol = newProtocol; newEligibility.ordering = pel_item.ordering; _documentDataModel.AddToProtocol_Eligibility(newEligibility); } And this is without the view: var pel_original = (from pel in _documentDataModel.Protocol_Eligibility where pel.Protocol.ID == id select pel); Protocol_Eligibility newEligibility; foreach (var pel_item in pel_original) { pel_item.EligibilityReference.Load(); newEligibility = new Protocol_Eligibility(); newEligibility.Eligibility = pel_item.Eligibility; newEligibility.Protocol = newProtocol; newEligibility.ordering = pel_item.ordering; _documentDataModel.AddToProtocol_Eligibility(newEligibility); }

    Read the article

  • C#/.NET Little Wonders: ConcurrentBag and BlockingCollection

    - by James Michael Hare
    In the first week of concurrent collections, began with a general introduction and discussed the ConcurrentStack<T> and ConcurrentQueue<T>.  The last post discussed the ConcurrentDictionary<T> .  Finally this week, we shall close with a discussion of the ConcurrentBag<T> and BlockingCollection<T>. For more of the "Little Wonders" posts, see C#/.NET Little Wonders: A Redux. Recap As you'll recall from the previous posts, the original collections were object-based containers that accomplished synchronization through a Synchronized member.  With the advent of .NET 2.0, the original collections were succeeded by the generic collections which are fully type-safe, but eschew automatic synchronization.  With .NET 4.0, a new breed of collections was born in the System.Collections.Concurrent namespace.  Of these, the final concurrent collection we will examine is the ConcurrentBag and a very useful wrapper class called the BlockingCollection. For some excellent information on the performance of the concurrent collections and how they perform compared to a traditional brute-force locking strategy, see this informative whitepaper by the Microsoft Parallel Computing Platform team here. ConcurrentBag<T> – Thread-safe unordered collection. Unlike the other concurrent collections, the ConcurrentBag<T> has no non-concurrent counterpart in the .NET collections libraries.  Items can be added and removed from a bag just like any other collection, but unlike the other collections, the items are not maintained in any order.  This makes the bag handy for those cases when all you care about is that the data be consumed eventually, without regard for order of consumption or even fairness – that is, it’s possible new items could be consumed before older items given the right circumstances for a period of time. So why would you ever want a container that can be unfair?  Well, to look at it another way, you can use a ConcurrentQueue and get the fairness, but it comes at a cost in that the ordering rules and synchronization required to maintain that ordering can affect scalability a bit.  Thus sometimes the bag is great when you want the fastest way to get the next item to process, and don’t care what item it is or how long its been waiting. The way that the ConcurrentBag works is to take advantage of the new ThreadLocal<T> type (new in System.Threading for .NET 4.0) so that each thread using the bag has a list local to just that thread.  This means that adding or removing to a thread-local list requires very low synchronization.  The problem comes in where a thread goes to consume an item but it’s local list is empty.  In this case the bag performs “work-stealing” where it will rob an item from another thread that has items in its list.  This requires a higher level of synchronization which adds a bit of overhead to the take operation. So, as you can imagine, this makes the ConcurrentBag good for situations where each thread both produces and consumes items from the bag, but it would be less-than-idea in situations where some threads are dedicated producers and the other threads are dedicated consumers because the work-stealing synchronization would outweigh the thread-local optimization for a thread taking its own items. Like the other concurrent collections, there are some curiosities to keep in mind: IsEmpty(), Count, ToArray(), and GetEnumerator() lock collection Each of these needs to take a snapshot of whole bag to determine if empty, thus they tend to be more expensive and cause Add() and Take() operations to block. ToArray() and GetEnumerator() are static snapshots Because it is based on a snapshot, will not show subsequent updates after snapshot. Add() is lightweight Since adding to the thread-local list, there is very little overhead on Add. TryTake() is lightweight if items in thread-local list As long as items are in the thread-local list, TryTake() is very lightweight, much more so than ConcurrentStack() and ConcurrentQueue(), however if the local thread list is empty, it must steal work from another thread, which is more expensive. Remember, a bag is not ideal for all situations, it is mainly ideal for situations where a process consumes an item and either decomposes it into more items to be processed, or handles the item partially and places it back to be processed again until some point when it will complete.  The main point is that the bag works best when each thread both takes and adds items. For example, we could create a totally contrived example where perhaps we want to see the largest power of a number before it crosses a certain threshold.  Yes, obviously we could easily do this with a log function, but bare with me while I use this contrived example for simplicity. So let’s say we have a work function that will take a Tuple out of a bag, this Tuple will contain two ints.  The first int is the original number, and the second int is the last multiple of that number.  So we could load our bag with the initial values (let’s say we want to know the last multiple of each of 2, 3, 5, and 7 under 100. 1: var bag = new ConcurrentBag<Tuple<int, int>> 2: { 3: Tuple.Create(2, 1), 4: Tuple.Create(3, 1), 5: Tuple.Create(5, 1), 6: Tuple.Create(7, 1) 7: }; Then we can create a method that given the bag, will take out an item, apply the multiplier again, 1: public static void FindHighestPowerUnder(ConcurrentBag<Tuple<int,int>> bag, int threshold) 2: { 3: Tuple<int,int> pair; 4:  5: // while there are items to take, this will prefer local first, then steal if no local 6: while (bag.TryTake(out pair)) 7: { 8: // look at next power 9: var result = Math.Pow(pair.Item1, pair.Item2 + 1); 10:  11: if (result < threshold) 12: { 13: // if smaller than threshold bump power by 1 14: bag.Add(Tuple.Create(pair.Item1, pair.Item2 + 1)); 15: } 16: else 17: { 18: // otherwise, we're done 19: Console.WriteLine("Highest power of {0} under {3} is {0}^{1} = {2}.", 20: pair.Item1, pair.Item2, Math.Pow(pair.Item1, pair.Item2), threshold); 21: } 22: } 23: } Now that we have this, we can load up this method as an Action into our Tasks and run it: 1: // create array of tasks, start all, wait for all 2: var tasks = new[] 3: { 4: new Task(() => FindHighestPowerUnder(bag, 100)), 5: new Task(() => FindHighestPowerUnder(bag, 100)), 6: }; 7:  8: Array.ForEach(tasks, t => t.Start()); 9:  10: Task.WaitAll(tasks); Totally contrived, I know, but keep in mind the main point!  When you have a thread or task that operates on an item, and then puts it back for further consumption – or decomposes an item into further sub-items to be processed – you should consider a ConcurrentBag as the thread-local lists will allow for quick processing.  However, if you need ordering or if your processes are dedicated producers or consumers, this collection is not ideal.  As with anything, you should performance test as your mileage will vary depending on your situation! BlockingCollection<T> – A producers & consumers pattern collection The BlockingCollection<T> can be treated like a collection in its own right, but in reality it adds a producers and consumers paradigm to any collection that implements the interface IProducerConsumerCollection<T>.  If you don’t specify one at the time of construction, it will use a ConcurrentQueue<T> as its underlying store. If you don’t want to use the ConcurrentQueue, the ConcurrentStack and ConcurrentBag also implement the interface (though ConcurrentDictionary does not).  In addition, you are of course free to create your own implementation of the interface. So, for those who don’t remember the producers and consumers classical computer-science problem, the gist of it is that you have one (or more) processes that are creating items (producers) and one (or more) processes that are consuming these items (consumers).  Now, the crux of the problem is that there is a bin (queue) where the produced items are placed, and typically that bin has a limited size.  Thus if a producer creates an item, but there is no space to store it, it must wait until an item is consumed.  Also if a consumer goes to consume an item and none exists, it must wait until an item is produced. The BlockingCollection makes it trivial to implement any standard producers/consumers process set by providing that “bin” where the items can be produced into and consumed from with the appropriate blocking operations.  In addition, you can specify whether the bin should have a limited size or can be (theoretically) unbounded, and you can specify timeouts on the blocking operations. As far as your choice of “bin”, for the most part the ConcurrentQueue is the right choice because it is fairly light and maximizes fairness by ordering items so that they are consumed in the same order they are produced.  You can use the concurrent bag or stack, of course, but your ordering would be random-ish in the case of the former and LIFO in the case of the latter. So let’s look at some of the methods of note in BlockingCollection: BoundedCapacity returns capacity of the “bin” If the bin is unbounded, the capacity is int.MaxValue. Count returns an internally-kept count of items This makes it O(1), but if you modify underlying collection directly (not recommended) it is unreliable. CompleteAdding() is used to cut off further adds. This sets IsAddingCompleted and begins to wind down consumers once empty. IsAddingCompleted is true when producers are “done”. Once you are done producing, should complete the add process to alert consumers. IsCompleted is true when producers are “done” and “bin” is empty. Once you mark the producers done, and all items removed, this will be true. Add() is a blocking add to collection. If bin is full, will wait till space frees up Take() is a blocking remove from collection. If bin is empty, will wait until item is produced or adding is completed. GetConsumingEnumerable() is used to iterate and consume items. Unlike the standard enumerator, this one consumes the items instead of iteration. TryAdd() attempts add but does not block completely If adding would block, returns false instead, can specify TimeSpan to wait before stopping. TryTake() attempts to take but does not block completely Like TryAdd(), if taking would block, returns false instead, can specify TimeSpan to wait. Note the use of CompleteAdding() to signal the BlockingCollection that nothing else should be added.  This means that any attempts to TryAdd() or Add() after marked completed will throw an InvalidOperationException.  In addition, once adding is complete you can still continue to TryTake() and Take() until the bin is empty, and then Take() will throw the InvalidOperationException and TryTake() will return false. So let’s create a simple program to try this out.  Let’s say that you have one process that will be producing items, but a slower consumer process that handles them.  This gives us a chance to peek inside what happens when the bin is bounded (by default, the bin is NOT bounded). 1: var bin = new BlockingCollection<int>(5); Now, we create a method to produce items: 1: public static void ProduceItems(BlockingCollection<int> bin, int numToProduce) 2: { 3: for (int i = 0; i < numToProduce; i++) 4: { 5: // try for 10 ms to add an item 6: while (!bin.TryAdd(i, TimeSpan.FromMilliseconds(10))) 7: { 8: Console.WriteLine("Bin is full, retrying..."); 9: } 10: } 11:  12: // once done producing, call CompleteAdding() 13: Console.WriteLine("Adding is completed."); 14: bin.CompleteAdding(); 15: } And one to consume them: 1: public static void ConsumeItems(BlockingCollection<int> bin) 2: { 3: // This will only be true if CompleteAdding() was called AND the bin is empty. 4: while (!bin.IsCompleted) 5: { 6: int item; 7:  8: if (!bin.TryTake(out item, TimeSpan.FromMilliseconds(10))) 9: { 10: Console.WriteLine("Bin is empty, retrying..."); 11: } 12: else 13: { 14: Console.WriteLine("Consuming item {0}.", item); 15: Thread.Sleep(TimeSpan.FromMilliseconds(20)); 16: } 17: } 18: } Then we can fire them off: 1: // create one producer and two consumers 2: var tasks = new[] 3: { 4: new Task(() => ProduceItems(bin, 20)), 5: new Task(() => ConsumeItems(bin)), 6: new Task(() => ConsumeItems(bin)), 7: }; 8:  9: Array.ForEach(tasks, t => t.Start()); 10:  11: Task.WaitAll(tasks); Notice that the producer is faster than the consumer, thus it should be hitting a full bin often and displaying the message after it times out on TryAdd(). 1: Consuming item 0. 2: Consuming item 1. 3: Bin is full, retrying... 4: Bin is full, retrying... 5: Consuming item 3. 6: Consuming item 2. 7: Bin is full, retrying... 8: Consuming item 4. 9: Consuming item 5. 10: Bin is full, retrying... 11: Consuming item 6. 12: Consuming item 7. 13: Bin is full, retrying... 14: Consuming item 8. 15: Consuming item 9. 16: Bin is full, retrying... 17: Consuming item 10. 18: Consuming item 11. 19: Bin is full, retrying... 20: Consuming item 12. 21: Consuming item 13. 22: Bin is full, retrying... 23: Bin is full, retrying... 24: Consuming item 14. 25: Adding is completed. 26: Consuming item 15. 27: Consuming item 16. 28: Consuming item 17. 29: Consuming item 19. 30: Consuming item 18. Also notice that once CompleteAdding() is called and the bin is empty, the IsCompleted property returns true, and the consumers will exit. Summary The ConcurrentBag is an interesting collection that can be used to optimize concurrency scenarios where tasks or threads both produce and consume items.  In this way, it will choose to consume its own work if available, and then steal if not.  However, in situations where you want fair consumption or ordering, or in situations where the producers and consumers are distinct processes, the bag is not optimal. The BlockingCollection is a great wrapper around all of the concurrent queue, stack, and bag that allows you to add producer and consumer semantics easily including waiting when the bin is full or empty. That’s the end of my dive into the concurrent collections.  I’d also strongly recommend, once again, you read this excellent Microsoft white paper that goes into much greater detail on the efficiencies you can gain using these collections judiciously (here). Tweet Technorati Tags: C#,.NET,Concurrent Collections,Little Wonders

    Read the article

  • django 1.1 beta issue

    - by ha22109
    Hello all, I m using django 1.1 beta.I m facing porblem in case of list_editable.First it was throughing exception saying need ordering in case of list_editable" then i added ordering in model but know it is giving me error.The code is working fine with django1.1 final. here is my code model.py class User(models.Model): advertiser = models.ForeignKey(WapUser,primary_key=True) status = models.CharField(max_length=20,choices=ADVERTISER_INVITE_STATUS,default='invited') tos_version = models.CharField(max_length=5) contact_email = models.EmailField(max_length=80) contact_phone = models.CharField(max_length=15) contact_mobile = models.CharField(max_length=15) contact_person = models.CharField(max_length=80) feedback=models.BooleanField(choices=boolean_choices,default=0) def __unicode__(self): return self.user.login class Meta: db_table = u'roi_advertiser_info' managed=False ordering=['feedback',] admin.py class UserAdmin(ReadOnlyAdminFields, admin.ModelAdmin): list_per_page = 15 fields = ['advertiser','contact_email','contact_phone','contact_mobile','contact_person'] list_display = ['advertiser','contact_email','contact_phone','contact_mobile','contact_person','status','feedback'] list_editable=['feedback'] readonly = ('advertiser',) search_fields = ['advertiser__login_id'] radio_fields={'approve_auto': admin.HORIZONTAL} list_filter=['status','feedback'] admin.site.register(User,UserADmin)

    Read the article

  • list editabale error

    - by ha22109
    Hello all, I m using django 1.1 beta.I m facing porblem in case of list_editable.First it was throughing exception saying need ordering in case of list_editable" then i added ordering in model but know it is giving me error.The code is working fine with django1.1 final. here is my code model.py class User(models.Model): advertiser = models.ForeignKey(WapUser,primary_key=True) status = models.CharField(max_length=20,choices=ADVERTISER_INVITE_STATUS,default='invited') tos_version = models.CharField(max_length=5) contact_email = models.EmailField(max_length=80) contact_phone = models.CharField(max_length=15) contact_mobile = models.CharField(max_length=15) contact_person = models.CharField(max_length=80) feedback=models.BooleanField(choices=boolean_choices,default=0) def __unicode__(self): return self.user.login class Meta: db_table = u'roi_advertiser_info' managed=False ordering=['feedback',] admin.py class UserAdmin(ReadOnlyAdminFields, admin.ModelAdmin): list_per_page = 15 fields = ['advertiser','contact_email','contact_phone','contact_mobile','contact_person'] list_display = ['advertiser','contact_email','contact_phone','contact_mobile','contact_person','status','feedback'] list_editable=['feedback'] readonly = ('advertiser',) search_fields = ['advertiser__login_id'] radio_fields={'approve_auto': admin.HORIZONTAL} list_filter=['status','feedback'] admin.site.register(User,UserADmin)

    Read the article

  • Optimizing ROW_NUMBER() in SQL Server

    - by BlueRaja
    We have a number of machines which record data into a database at sporadic intervals. For each record, I'd like to obtain the time period between this recording and the previous recording. I can do this using ROW_NUMBER as follows: WITH TempTable AS ( SELECT *, ROW_NUMBER() OVER (PARTITION BY Machine_ID ORDER BY Date_Time) AS Ordering FROM dbo.DataTable ) SELECT [Current].*, Previous.Date_Time AS PreviousDateTime FROM TempTable AS [Current] INNER JOIN TempTable AS Previous ON [Current].Machine_ID = Previous.Machine_ID AND Previous.Ordering = [Current].Ordering + 1 The problem is, it goes really slow (several minutes on a table with about 10k entries) - I tried creating separate indicies on Machine_ID and Date_Time, and a single joined-index, but nothing helps. Is there anyway to rewrite this query to go faster?

    Read the article

  • Algorithm - combine multiple lists, resulting in unique list and retaining order

    - by hitch
    I want to combine multiple lists of items into a single list, retaining the overall order requirements. i.e.: 1: A C E 2: D E 3: B A D result: B A C D E above, starting with list 1, we have ACE, we then know that D must come before E, and from list 3, we know that B must come before A, and D must come after B and A. If there are conflicting orderings, the first ordering should be used. i.e. 1: A C E 2: B D E 3: F D result: A C B D E F 3 conflicts with 2, therefore requirements for 2 will be used. If ordering requirements mean an item must come before or after another, it doesn't matter if it comes immediately before or after, or at the start or end of the list, as long as overall ordering is maintained. This is being developed using VB.Net, so a LINQy solution (or any .Net solution) would be nice - otherwise pointers for an approach would be good.

    Read the article

  • C# 4.0: Covariance And Contravariance In Generics

    - by Paulo Morgado
    C# 4.0 (and .NET 4.0) introduced covariance and contravariance to generic interfaces and delegates. But what is this variance thing? According to Wikipedia, in multilinear algebra and tensor analysis, covariance and contravariance describe how the quantitative description of certain geometrical or physical entities changes when passing from one coordinate system to another.(*) But what does this have to do with C# or .NET? In type theory, a the type T is greater (>) than type S if S is a subtype (derives from) T, which means that there is a quantitative description for types in a type hierarchy. So, how does covariance and contravariance apply to C# (and .NET) generic types? In C# (and .NET), variance applies to generic type parameters and not to the resulting generic type. A generic type parameter is: covariant if the ordering of the generic types follows the ordering of the generic type parameters: Generic<T> = Generic<S> for T = S. contravariant if the ordering of the generic types is reversed from the ordering of the generic type parameters: Generic<T> = Generic<S> for T = S. invariant if neither of the above apply. If this definition is applied to arrays, we can see that arrays have always been covariant because this is valid code: object[] objectArray = new string[] { "string 1", "string 2" }; objectArray[0] = "string 3"; objectArray[1] = new object(); However, when we try to run this code, the second assignment will throw an ArrayTypeMismatchException. Although the compiler was fooled into thinking this was valid code because an object is being assigned to an element of an array of object, at run time, there is always a type check to guarantee that the runtime type of the definition of the elements of the array is greater or equal to the instance being assigned to the element. In the above example, because the runtime type of the array is array of string, the first assignment of array elements is valid because string = string and the second is invalid because string = object. This leads to the conclusion that, although arrays have always been covariant, they are not safely covariant – code that compiles is not guaranteed to run without errors. In C#, the way to define that a generic type parameter as covariant is using the out generic modifier: public interface IEnumerable<out T> { IEnumerator<T> GetEnumerator(); } public interface IEnumerator<out T> { T Current { get; } bool MoveNext(); } Notice the convenient use the pre-existing out keyword. Besides the benefit of not having to remember a new hypothetic covariant keyword, out is easier to remember because it defines that the generic type parameter can only appear in output positions — read-only properties and method return values. In a similar way, the way to define a type parameter as contravariant is using the in generic modifier: public interface IComparer<in T> { int Compare(T x, T y); } Once again, the use of the pre-existing in keyword makes it easier to remember that the generic type parameter can only be used in input positions — write-only properties and method non ref and non out parameters. Because covariance and contravariance apply only to the generic type parameters, a generic type definition can have both covariant and contravariant generic type parameters in its definition: public delegate TResult Func<in T, out TResult>(T arg); A generic type parameter that is not marked covariant (out) or contravariant (in) is invariant. All the types in the .NET Framework where variance could be applied to its generic type parameters have been modified to take advantage of this new feature. In summary, the rules for variance in C# (and .NET) are: Variance in type parameters are restricted to generic interface and generic delegate types. A generic interface or generic delegate type can have both covariant and contravariant type parameters. Variance applies only to reference types; if you specify a value type for a variant type parameter, that type parameter is invariant for the resulting constructed type. Variance does not apply to delegate combination. That is, given two delegates of types Action<Derived> and Action<Base>, you cannot combine the second delegate with the first although the result would be type safe. Variance allows the second delegate to be assigned to a variable of type Action<Derived>, but delegates can combine only if their types match exactly. If you want to learn more about variance in C# (and .NET), you can always read: Covariance and Contravariance in Generics — MSDN Library Exact rules for variance validity — Eric Lippert Events get a little overhaul in C# 4, Afterward: Effective Events — Chris Burrows Note: Because variance is a feature of .NET 4.0 and not only of C# 4.0, all this also applies to Visual Basic 10.

    Read the article

  • Move Window Buttons Back to the Right in Ubuntu 10.04

    - by Trevor Bekolay
    One of the more controversial changes in the Ubuntu 10.04 beta is the Mac OS-inspired change to have window buttons on the left side. We’ll show you how to move the buttons back to the right. Before While the change may or may not persist through to the April 29 release of Ubuntu 10.04, in the beta version the maximize, minimize, and close buttons appear in the top left of a window. How to move the window buttons The window button locations are dictated by a configuration file. We’ll use the graphical program gconf-editor to change this configuration file. Press Alt+F2 to bring up the Run Application dialog box, enter “gconf-editor” in the text field, and click on Run. The Configuration Editor should pop up. The key that we want to edit is in apps/metacity/general. Click on the + button next to the “apps” folder, then beside “metacity” in the list of folders expanded for apps, and then click on the “general” folder. The button layout can be changed by changing the “button_layout” key. Double-click button_layout to edit it. Change the text in the Value text field to: menu:maximize,minimize,close Click OK and the change will occur immediately, changing the location of the window buttons in the Configuration Editor. Note that this ordering of the window buttons is slightly different than the typical order; in previous versions of Ubuntu and in Windows, the minimize button is to the left of the maximize button. You can change the button_layout string to reflect that ordering, but using the default Ubuntu 10.04 theme, it looks a bit strange. If you plan to change the theme, or even just the graphics used for the window buttons, then this ordering may be more natural to you. After After this change, all of your windows will have the maximize, minimize, and close buttons on the right. What do you think of Ubuntu 10.04’s visual change? Let us know in the comments! Similar Articles Productive Geek Tips Move a Window Without Clicking the Titlebar in UbuntuBring Misplaced Off-Screen Windows Back to Your Desktop (Keyboard Trick)Keep the Display From Turning Off on UbuntuPut Close/Maximize/Minimize Buttons on the Left in UbuntuAllow Remote Control To Your Desktop On Ubuntu TouchFreeze Alternative in AutoHotkey The Icy Undertow Desktop Windows Home Server – Backup to LAN The Clear & Clean Desktop Use This Bookmarklet to Easily Get Albums Use AutoHotkey to Assign a Hotkey to a Specific Window Latest Software Reviews Tinyhacker Random Tips Revo Uninstaller Pro Registry Mechanic 9 for Windows PC Tools Internet Security Suite 2010 PCmover Professional SpeedyFox Claims to Speed up your Firefox Beware Hover Kitties Test Drive Mobile Phones Online With TryPhone Ben & Jerry’s Free Cone Day, 3/23/10 New Stinger from McAfee Helps Remove ‘FakeAlert’ Threats Google Apps Marketplace: Tools & Services For Google Apps Users

    Read the article

  • How does one find out which application is associated with an indicator icon?

    - by Amos Annoy
    It is trivial to do this in Ubuntu 10.04. The question is specific to Ubuntu 12.04. some pertinent references (src: answer to What is the difference between indicators and a system tray?: Here is the documentation for indicators: Application indicators | Ubuntu App Developer libindicate Reference Manual libappindicator Reference Manual also DesktopExperienceTeam/ApplicationIndicators - Ubuntu Wiki ref: How can the application that makes an indicator icon be identified? bookmark: How does one find out which application is associated with an indicator icon in Ubuntu 12.04? is a serious question for reasons & problems outlined below and for which a significant investment has been made and is necessary for remedial purposes. reviewing refs. to find an orchestrated resolution ... (an indicator ap. indicator maybe needed) This has nothing to do (does it?) with right click. How can an indicator's icon in Ubuntu 12.04 be matched with the program responsible for it's manifestation on the top panel? A list of running applications can include all processes using System Monitor. How is the correct matching process found for an indicator? How are the sub-indicator applications identified? These are the aps associated with the components of an indicators drop-down menu. (This was to be a separate question and quite naturally follows up the progression. It is included here as it is obvious there is no provisioning to track down offending either sub or indicator aps. easily.) (The examination of SM points out a rather poignant factor in the faster battery depletion and shortened run time - the ambient quiescent CPU rate in 12.04 is now well over 20% when previously, in 10.04, it was well under 10%, between 5% and 7%! - the huge inordinate cpu overhead originates from Xorg and compiz - after booting the system, only SM is run and All Processes are selected, sorting on %CPU - switching between Resources and Processes profiles the execution overhead problem - running another ap like gedit "Text Editor" briefly gives it CPU priority - going back to S&M several aps. are at the top of the list in order: gnome-system-monitor as expected, then: Xorg, compiz, unity-panel-service, hud-service, with dbus-daemon and kworker/x:y's mixed in with some expected daemons and background tasks like nm-applet - not only do Xorg and compiz require excessive CPU time but their entourage has to come along too! further exacerbating the problem - our compute bound tasks no longer work effectively in the field - reduced battery life, reduced CPU time for custom ap.s etc. - and all this precipitated from an examination of what is going on with the battery ap. indicator - this was and is not a flippant, rhetorical or idle musing but has consequences for the credible deployment of 12.04 to reduce the negative impact of its overhead in a production environment) (I have a problem with the battery indicator - it sometimes has % and other times hh:mm - it is necessary to know the ap. & v. to get more info on controlling same. ditto: There are issues with other indicator aps.: NM vs. iwlist/iwconfig conflict, BT ap. vs RF switch, Battery ap. w/ no suspend/sleep for poor battery runtime, ... the list goes on) Details from: How can I find Application Indicator ID's? suggests looking at: file:///usr/share/indicator-application/ordering-override.keyfile [Ordering Index Overrides] nm-applet=1 gnome-power-manager=2 ibus=3 gst-keyboard-xkb=4 gsd-keyboard-xkb=5 which solves the battery ap. identification, and presumably nm is NetworkManager for the rf icon, but the envelope, blue tooth and speaker indicator aps. are still a mystery. (Also, the ordering is not correlated.) Mind you, it was simple in the past to simply right click to get the About option to find the ap. & v. info. browsing around and about: file:///usr/share/indicator-application/ordering-override.keyfile examined: file:///usr/share/indicators file:///usr/share/indicators/messages/applications/ ... perhaps?/presumably? the information sought may be buried in file:///usr/share/indicators A reference in the comments was given to: What is the difference between indicators and a system tray? quoting from that source ... Unfortunately desktop indicators are not well documented yet: I couldn't find any specification doc ... Well ... the actual document https://wiki.ubuntu.com/DesktopExperienceTeam/ApplicationIndicators#Summary does not help much but it's existential information provides considerable insight ...

    Read the article

  • What Design Pattern is seperating transform converters

    - by RevMoon
    For converting a Java object model into XML I am using the following design: For different types of objects (e.g. primitive types, collections, null, etc.) I define each its own converter, which acts appropriate with respect to the given type. This way it can easily extended without adding code to a huge if-else-then construct. The converters are chosen by a method which tests whether the object is convertable at all and by using a priority ordering. The priority ordering is important so let's say a List is not converted by the POJO converter, even though it is convertable as such it would be more appropriate to use the collection converter. What design pattern is that? I can only think of a similarity to the command pattern.

    Read the article

  • What Design Pattern is separating transform converters

    - by RevMoon
    For converting a Java object model into XML I am using the following design: For different types of objects (e.g. primitive types, collections, null, etc.) I define each its own converter, which acts appropriate with respect to the given type. This way it can easily extended without adding code to a huge if-else-then construct. The converters are chosen by a method which tests whether the object is convertable at all and by using a priority ordering. The priority ordering is important so let's say a List is not converted by the POJO converter, even though it is convertable as such it would be more appropriate to use the collection converter. What design pattern is that? I can only think of a similarity to the command pattern.

    Read the article

  • See how one goal leads to another (i.e., how often downloading the demo leads to purchasing)

    - by s3cur3
    I have two goals set up for my site in Google Analytics: one to download the demo of my software, and the other to buy the full version. I'm having trouble getting statistics on how often the download leads to a purchase. This seems like something the multi-channel funnel is perfect for (as in this question)---I expect that it takes multiple visits for the demo download to convert into an order. However, I can't seem to narrow down the multi-channel funnel data into only those that have one goal followed by another. If it matters, I'm tracking the ordering action as a page view (i.e., someone views the "Thanks for ordering" page) and the demo download as an action (which fires when someone clicks the "download" button). Thanks! - Tyler

    Read the article

< Previous Page | 4 5 6 7 8 9 10 11 12 13 14 15  | Next Page >