Search Results

Search found 10312 results on 413 pages for 'compiler bug'.

Page 17/413 | < Previous Page | 13 14 15 16 17 18 19 20 21 22 23 24  | Next Page >

  • Bug in CF9: values for unique struct keys referenced and overwritten by other keys.

    - by Gin Doe
    We've run into a serious issue with CF9 wherein values for certain struct keys can be referenced by other keys, despite those other keys never being set. See the following examples: Edit: Looks like it isn't just something our servers ate. This is Adobe bug-track ticket 81884: http://cfbugs.adobe.com/cfbugreport/flexbugui/cfbugtracker/main.html#bugId=81884. <cfset a = { AO = "foo" } /> <cfset b = { AO = "foo", B0 = "bar" } /> <cfoutput> The following should throw an error. Instead both keys refer to the same value. <br />Struct a: <cfdump var="#a#" /> <br />a.AO: #a.AO# <br />a.B0: #a.B0# <hr /> The following should show a struct with 2 distinct keys and values. Instead it contains a single key, "AO", with a value of "bar". <br />Struct b: <cfdump var="#b#" /> This is obviously a complete show-stopper for us. I'd be curious to know if anyone has encountered this or can reproduce this in their environment. For us, it happens 100% of the time on Apache/CF9 running on Linux, both RH4 and RH5. We're using the default JRun install on Java 1.6.0_14. To see the extent of the problem, we ran a quick loop to find other naming sequences that are affected and found hundreds of matches for 2 letter key names. A similar loop found more conflicts in 3 letter names. <cfoutput>Testing a range of affected key combinations. This found hundreds of cases on our platform. Aborting after 50 here.</cfoutput> <cfscript> teststring = "ABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789"; stringlen = len(teststring); matchesfound = 0; matches = ""; for (i1 = 1; i1 <= stringlen; i1++) { symbol1 = mid(teststring, i1, 1); for (i2 = 1; i2 <= stringlen; i2++) { teststruct = structnew(); symbol2 = mid(teststring, i2, 1); symbolwhole = symbol1 & symbol2; teststruct[ symbolwhole ] = "a string"; for (q1 = 1; q1 <= stringlen; q1++) { innersymbol1 = mid(teststring, q1, 1); for (q2 = 1; q2 <= stringlen; q2++) { innersymbol2 = mid(teststring, q2, 1); innersymbolwhole = innersymbol1 & innersymbol2; if ((i1 != q1 || i2 != q2) && structkeyexists(teststruct, innersymbolwhole)) { // another affected pair of keys! writeoutput ("<br />#symbolwhole# = #innersymbolwhole#"); if (matchesfound++ > 50) { // we've seen enough abort; } } } } } } </cfscript> And edit again: This doesn't just affect struct keys but names in the variables scope as well. At least the variables scope has the presence of mind to throw an error, "can't load a null": <cfset test_b0 = "foo" /> <cfset test_ao = "bar" /> <cfoutput> test_b0: #test_b0# <br />test_ao: #test_ao# </cfoutput>

    Read the article

  • Speedup of fixing an openssl bug with 8192 bit key [on hold]

    - by rubo77
    This is related to this Bug-Report https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=747453 OpenSSL contains a set of arbitrary limitations on the size of accepted key parameters that make unrelated software fail to establish secure connections. The problem was found while debugging a XMPP s2s connection issue where two servers with long certificate keys (8192 Bit RSA) failed to establish a secure connection because OpenSSL rejected the handshake. This seems to be a small problem to be fixed but although there is an easy patch available to fix the issue in that bug report, no reactions are noticed so far.. The last patch that broke the 2048 barrier took 2 years to be implemented and only resulted in an increase to 4096bit, which seems to be a bad joke. Where would we have to report this to speed up the implementation for such an issue?

    Read the article

  • Can I make the compiler and the CLR ignore non implemented interfaces on my types?

    - by Thiado de Arruda
    I would like to define a type implementing a certain interface, however I would only implement it in a proxy at runtime. I can see two obstacles in this scenario : 1-Make the compiler ignore non implemented interfaces. 2-Make the CLR ignore(or at least delay) the TypeLoadException with the following description : "Method SOMEMETHOD in type SOMETYPE from assembly SOMEASSEMBLY does not have an implementation." Is something like this possible?

    Read the article

  • Make compiler copy characters using movsd

    - by Suma
    I would like to copy a relatively short sequence of memory (less than 1 KB, typically 2-200 bytes) in a time critical function. The best code for this on CPU side seems to be rep movsd. However I somehow cannot make my compiler to generate this code. I hoped (and I vaguely remember seeing so) using memcpy would do this using compiler built-in instrinsic, but based on disassembly and debugging it seems compiler is using call to memcpy/memmove library implementation instead. I also hoped the compiler might be smart enough to recognize following loop and use rep movsd on its own, but it seems it does not. char *dst; const char *src; // ... for (int r=size; --r>=0; ) *dst++ = *src++; Is there some way to make the Visual Studio compiler to generate rep movsd sequence other than using inline assembly?

    Read the article

  • Why will this code compile using ifort compiler and not when using gfortran compiler? Help!

    - by CuriousCompiler
    I'm rewriting some code to make a program compile with the gfortran compiler as opposed to ifort compiler I usually use. The code follows: _Subroutine SlideBits (WORD, BITS, ADDR) Implicit None Integer(4) WORD Integer(4) BITS Integer(4) ADDR Integer(4) ADDR1 ADDR1 = 32 - ADDR WORD = (WORD .And. (.Not.ISHFT(1,ADDR1))) .Or. ISHFT(BITS,ADDR1) End_ When I compile the above code using the gfortran compiler, I recieve this error: WORD = (WORD .And. (.Not.ISHFT(1,ADDR1))) .Or. ISHFT(BITS,ADDR1) Error: Operand of .NOT. operator at (1) is INTEGER(4) All three of the variables coming into the subroutine are integers. I've looked around a bit and the gfortran wiki states that the gfortran compiler should be able to handle logical statments being applied to integer values. Several other sites I've visited either quote from the gnu wiki or agree with it. This is the first time I've seen this error as the Intel Fortran compiler (ifort) I normally use compiles cleanly.

    Read the article

  • How to fix Monogame WP8 Touch Position bug?

    - by Moses Aprico
    Normally below code will result in X:Infinity, Y:Infinity TouchCollection touchState = TouchPanel.GetState(); foreach (TouchLocation t in touchState) { if (t.State == TouchLocationState.Pressed) { vb.ButtonTouched((int)t.Position.X, (int)t.Position.Y); } } Then, I followed this https://github.com/mono/MonoGame/issues/1046 and added below code at the first line in update method. (I still don't know how it's worked, but it fixed the problem) if (_firstUpdate) { typeof(Microsoft.Xna.Framework.Input.Touch.TouchPanel).GetField("_touchScale",System.Reflection.BindingFlags.NonPublic | System.Reflection.BindingFlags.Static).SetValue(null, Vector2.One); _firstUpdate = false; } And then, when I randomly testing something, there are several area that won't read the user touch. The tile with the purple dude is the area which won't receive user input (It don't even detect "Pressed", the TouchCollection.Count = 0) Any idea how to fix this? UPDATE 1 : The second attempt in recompiling The difference is weird. Dunno why the consistent clickable area is just 2/3 area to the left UPDATE 2 : After trying to rotate to landscape and back to portrait to randomly testing, then the outcome become :

    Read the article

  • Oracle EBS R12.1.1 system09.dbf file corruption Bug

    - by longchun.zhu
    ??????,??????????????????,???? ?????????????.. ???????????,??????,???????????? After Installing or Upgrading Perform the following steps after installing or upgrading to Release 12.1.1 and before allowing users to access the system. Manually fix database dbf file If you installed 12.1.1 with a startCD of 12.1.1.9 or earlier (see Oracle Applications Release Notes, Release 12.1.1 My Oracle Support Document 798258.1), you must run the following sql commands to fix a particular corrupted dbf file: $ sqlplus/nolog sql connect / as sysdba sql alter database datafile '[full path of system09.dbf]' resize 1000M; sql alter database datafile '[full path of system09.dbf]' resize 1500M;

    Read the article

  • HUGE EF4 Inheritance Bug

    - by djsolid
    Well maybe not for everyone but for me is definitely really important. That is why I get straight into the point. We have the following model: Which maps to the following database: We are using EF4.0 and we want to load all Burgers including BurgerDetails. So we write the following query: But it fails. The error is : “The ResultType of the specified expression is not compatible with the required type. The expression ResultType is 'Transient.reference[SampleEFDBModel.Food]' but the required type is 'Transient.reference[SampleEFDBModel.Burger]'.Parameter name: arguments[0]”   So in the new version of EF there is no way to eager load data through Navigation Properties with 1-1 relationships defined in subclasses. Here is the relevant Microsoft Connect Issue. It is described through an other example but the result is the same.  Please if you think this is important vote up on Microsoft Connect.   EF 4.0 has many improvements. I am using it since v1 in large-scale projects and this version is faster,produces cleaner sql, more reliable and can be used for complicated business scenarios. That is why I believe this issue should be solved as soon as possible. I understand that release cycles are slow but I am hoping atleast for a hotfix. I also have uploaded the example project so you can test it. Download it from here. If anyone has found any workarounds please post it in the comments section. Thanks!

    Read the article

  • Screen (command-line program) bug?

    - by VioletCrime
    fired up my Minecraft server again after about a year off. My server used to run 11.04, which has since been upgraded to 12.04; I lost my management scripts in the upgrade (thought I'd backed up the user's home directory), but whatever, I enjoy developing stuff like that anyways. However, this time around, I'm running into issues. I start the Minecraft server using a detached screen, however the script is unable to 'stuff' commands into the screen instance until I attach to the screen then detach again? Once I do that, I can stuff anything I want into the server's terminal using the -X option, until I stop/start the server again, then I have to reattach and detach in order to restore functionality? Here's the manager script: #!/bin/bash #Name of the screen housing the server (set with the -S flag at startup) SCREENNAME=minecraft1 #Name of the folder housing the Minecraft world FOLDERNAME=world1 startServer (){ if screen -list | grep "$SCREENNAME" > /dev/null then echo "Cannot start Minecraft server; it is already running!" else screen -dmS $SCREENNAME java -Xmx1024M -Xms1024M -jar minecraft_server.jar sleep 2 if screen -list | grep "$SCREENNAME" > /dev/null then echo "Minecraft server started; happy mining!" else echo "ERROR: Minecraft server failed to start!" fi fi; } stopServer (){ if screen -list | grep "$SCREENNAME" > /dev/null then echo "Server is running. Giving a 1-minute warning." screen -S $SCREENNAME -X stuff "/say Shutting down (halt) in one minute." screen -S $SCREENNAME -X stuff $'\015' sleep 45 screen -S $SCREENNAME -X stuff "/say Shutting down (halt) in 15 seconds." screen -S $SCREENNAME -X stuff $'\015' sleep 15 screen -S $SCREENNAME -X stuff "/stop" screen -S $SCREENNAME -X stuff $'\015' else echo "Server is not running; nothing to stop." fi; } stopServerNow (){ if screen -list | grep "$SCREENNAME" > /dev/null then echo "Server is running. Giving a 5-second warning." screen -S $SCREENNAME -X stuff "/say EMERGENCY SHUTDOWN! 5 seconds to halt." screen -S $SCREENNAME -X stuff $'\015' sleep 5 screen -S $SCREENNAME -X stuff "/stop" screen -S $SCREENNAME -X stuff $'\015' else echo "Server is not running; nothing to stop." fi; } restartServer (){ if screen -list | grep "$SCREENNAME" > /dev/null then echo "Server is running. Giving a 1-minute warning." screen -S $SCREENNAME -X stuff "/say Shutting down (restart) in one minute." screen -S $SCREENNAME -X stuff $'\015' sleep 45 screen -S $SCREENNAME -X stuff "/say Shutting down (restart) in 15 seconds." screen -S $SCREENNAME -X stuff $'\015' sleep 15 screen -S $SCREENNAME -X stuff "/stop" screen -S $SCREENNAME -X stuff $'\015' sleep 2 startServer else echo "Cannot restart server: it isn't running." fi; } #In order for this function to work, a directory 'backup/$FOLDERNAME' must exist in the same #directory that '$FOLDERNAME' resides backupWorld (){ if screen -list | grep "$SCREENNAME" > /dev/null then echo "Server is running. Giving a 1-minute warning." screen -S $SCREENNAME -X stuff "/say Shutting down (backup) in one minute." screen -S $SCREENNAME -X stuff $'\015' screen -S $SCREENNAME -X stuff "/say Server should be down for no more than a few seconds." screen -S $SCREENNAME -X stuff $'\015' sleep 45 screen -S $SCREENNAME -X stuff "/say Shutting down for backup in 15 seconds." screen -S $SCREENNAME -X stuff $'\015' sleep 15 screen -S $SCREENNAME -X stuff "/stop" screen -S $SCREENNAME -X stuff $'\015' fi sleep 2 if screen -list | grep $SCREENNAME > /dev/null then echo "Server is still running? Error." else cd .. tar -czvf backup0.tar.gz $FOLDERNAME mv backup0.tar.gz backup/$FOLDERNAME cd backup/$FOLDERNAME rm backup10.tar.gz mv backup9.tar.gz backup10.tar.gz mv backup8.tar.gz backup9.tar.gz mv backup7.tar.gz backup8.tar.gz mv backup6.tar.gz backup7.tar.gz mv backup5.tar.gz backup6.tar.gz mv backup4.tar.gz backup5.tar.gz mv backup3.tar.gz backup4.tar.gz mv backup2.tar.gz backup3.tar.gz mv backup1.tar.gz backup2.tar.gz mv backup0.tar.gz backup1.tar.gz cd ../../$FOLDERNAME screen -dmS $SCREENNAME java -Xmx1024M -Xms1024M -jar minecraft_server.jar; sleep 2 if screen -list | grep "$SCREENNAME" > /dev/null then echo "Minecraft server restarted; happy mining!" else echo "ERROR: Minecraft server failed to start!" fi fi; } printCommands (){ echo echo "$0 usage:" echo echo "Start : Starts the server on a detached screen." echo "Stop : Stop the server; includes a 1-minute warning." echo "StopNOW : Stops the server with only a 5-second warning." echo "Restart : Stops the server and starts the server again." echo "Backup : Stops the server (1 min), backs up the world, and restarts." echo "Help : Display this message." } #Forces case-insensitive string comparisons shopt -s nocasematch #Primary 'Switch' if [[ $1 = "start" ]] then startServer elif [[ $1 = "stop" ]] then stopServer elif [[ $1 = "stopnow" ]] then stopServerNow elif [[ $1 = "backup" ]] then backupWorld elif [[ $1 = "restart" ]] then restartServer else printCommands fi

    Read the article

  • Find The Bug

    - by Alois Kraus
    What does this code print and why?             HashSet<int> set = new HashSet<int>();             int[] data = new int[] { 1, 2, 1, 2 };             var unique = from i in data                          where set.Add(i)                          select i;   // Compiles to: var unique = Enumerable.Where(data, (i) => set.Add(i));             foreach (var i in unique)             {                 Console.WriteLine("First: {0}", i);             }               foreach (var i in unique)             {                 Console.WriteLine("Second: {0}", i);             }   The output is: First: 1 First: 2 Why is there no output of the second loop? The reason is that LINQ does not cache the results of the collection but it does recalculate the contents for every new enumeration again. Since I have used state (the Hashset does decide which entries are part of the output) I do arrive with an empty sequence since Add of the Hashset will return false for all values I have already passed in leaving nothing to return a second time. The solution is quite simple: Use the Distinct extension method or cache the results by calling .ToList() or ToArray() for the result of the LINQ query. Lession Learned: Do never forget to think about state in Where clauses!

    Read the article

  • Bug! Slow Sums and Averages

    - by Paul White
    It’s a curious thing about SQL that the SUM or AVG of no items is not zero, it’s NULL. In this post, you’ll see how this means your SUM and AVG calculations might run at half speed, or worse. As with most of my blog entries though, today’s instalment is not so much about the result, but the journey we take to get there. Before we get started on that, I just want to mention that there’s a problem with the Google Reader feed for this blog, so those of you that use that will have missed two recent entries: Seeking Without Indexes and Advanced TSQL Tuning: Why Internals Knowledge Matters. Accessing the site directly always works of course :) Ok, on to today’s story. Take a look at this query:...(read more)

    Read the article

  • Fixing Chrome&rsquo;s AJAX Request Caching Bug

    - by Steve Wilkes
    I recently had to make a set of web pages restore their state when the user arrived on them after clicking the browser’s back button. The pages in question had various content loaded in response to user actions, which meant I had to manually get them back into a valid state after the page loaded. I got hold of the page’s data in a JavaScript ViewModel using a JQuery ajax call, then iterated over the properties, filling in the fields as I went. I built in the ability to describe dependencies between inputs to make sure fields were filled in in the correct order and at the correct time, and that all worked nicely. To make sure the browser didn’t cache the AJAX call results I used the JQuery’s cache: false option, and ASP.NET MVC’s OutputCache attribute for good measure. That all worked perfectly… except in Chrome. Chrome insisted on retrieving the data from its cache. cache: false adds a random query string parameter to make the browser think it’s a unique request – it made no difference. I made the AJAX call a POST – it made no difference. Eventually what I had to do was add a random token to the URL (not the query string) and use MVC routing to deliver the request to the correct action. The project had a single Controller for all AJAX requests, so this route: routes.MapRoute( name: "NonCachedAjaxActions", url: "AjaxCalls/{cacheDisablingToken}/{action}", defaults: new { controller = "AjaxCalls" }, constraints: new { cacheDisablingToken = "[0-9]+" }); …and this amendment to the ajax call: function loadPageData(url) { // Insert a timestamp before the URL's action segment: var indexOfFinalUrlSeparator = url.lastIndexOf("/"); var uniqueUrl = url.substring(0, indexOfFinalUrlSeparator) + new Date().getTime() + "/" + url.substring(indexOfFinalUrlSeparator); // Call the now-unique action URL: $.ajax(uniqueUrl, { cache: false, success: completePageDataLoad }); } …did the trick.

    Read the article

  • MySql Connector/NET 6.7.4 GA has been released

    - by fernando
    MySQL Connector/Net 6.7.4, a new version of the all-managed .NET driver for MySQL has been released.  This is the GA, is feature complete. It is recommended for production environments.  It is appropriate for use with MySQL server versions 5.0-5.7.  It is now available in source and binary form from http://dev.mysql.com/downloads/connector/net/#downloads and mirror sites (note that not all mirror sites may be up to date at this point-if you can't find this version on some mirror, please try again later or choose another download site.) The 6.7 version of MySQL Connector/Net brings the following new features: -  WinRT Connector. -  Load Balancing support. -  Entity Framework 5.0 support. -  Memcached support for Innodb Memcached plugin. -  This version also splits the product in two: from now on, starting version 6.7, Connector/NET will include only the former Connector/NET ADO.NET driver, Entity Framework and ASP.NET providers (Core libraries of MySql.Data, MySql.Data.Entity & MySql.Web). While all the former product Visual Studio integration (Design support, Intellisense, Debugger) are available as part of MySql Windows Installer under the name "MySql for Visual Studio".  WinRT Connector  ------------------------------------------- Now you can write MySql data access apps in Windows Runtime (aka Store Apps) using the familiar API of Connector/NET for .NET.  Load Balancing Support  -------------------------------------------  Now you can setup a Replication or Cluster configuration in the backend, and Connector/NET will balance the load of queries among all servers making up the backend topology.  Entity Framework 5.0  -------------------------------------------  Connector/NET is now compatible with EF 5, including special features of EF 5 like spatial types.  Memcached  -------------------------------------------  Just setup Innodb memcached plugin and use Connector/NET new APIs to establish a client to MySql 5.6 server's memcached daemon.  Bug fixes included in this release: - Fix for Entity Framework when inserts data having Identity columns (Oracle bug #16494585). - Fix for Connector/NET cannot read data from a MySql table using UTF-16/UTF-32 (MySql bug #69169, Oracle bug #16776818). - Fix for Malformed query in Entity Framework when eager loading due to multiple projections (MySql bug #67183, Oracle bug #16872852). - Fix for database objects with 'dbo' prefix when using automatic migrations in Entity Framework 5.0 (Oracle bug #16909439). - Fix for bug IIS application pool reset worker process causes website to crash (Oracle bug #16909237, Mysql Bug #67665). - Fix for bug Error in LINQ to Entities query when using Distinct().Count() (MySql Bug #68513, Oracle bug #16950146). - Fix for occasionally return no data when socket connection is slow, interrupted or delayed (MySql bug #69039, Oracle bug #16950212). - Fix for ConstraintException when filling a datatable (MySql bug #65065, Oracle bug #16952323). - Fix for Data Provider is not found after uninstalling Mysql for visual studio (Oracle bug #16973456). - Fix for nested sql generated for LINQ to Entities query with Take and Order by (MySql bug #65723, Oracle bug #16973939). The documentation is available at http://dev.mysql.com/doc/refman/5.7/en/connector-net.html  Enjoy and thanks for the support!  --  Fernando Gonzalez Sanchez | Software Engineer |  Oracle MySQL Windows Experience Team, Connector/NET  Guadalajara | Jalisco | Mexico 

    Read the article

  • Bug unsubscribing from Ubuntu One Mobile

    - by rhino
    Hi guys, I have an Ubuntu One Mobile subscription, which I can see in my subscriptions page: one.ubuntu.com/account/subscription/756082 I no longer need my Ubuntu One Mobile subscription, so click the link to cancel the Mobile service subscription: one.ubuntu.com/account/cancel/756082/ Then confirm that request to cancel: one.ubuntu.com/account/cancel/756082/confirm/ But the process ends there showing a "Something has gone wrong page", and my subscription remains active :( The same problem occurred when I attempted the same a few weeks back, so not a temporary problem I'm thinking. Any input gratefully received. I would like to report this problem directly to the maintainer of this part of the Ubuntu site but cannot see how to do that.

    Read the article

  • Understanding hand written lexers

    - by Cole Johnson
    I am going to make a compiler for C (C99; I own the standards PDF), written in C (go figure) and looking up on how compilers work on Wikipedia has told me a lot. However, after reading up on lexers has confused me. The Wikipedia page states that: the GNU Compiler Collection (gcc) uses hand-written lexers I have tried googling what a hand written lexer and have come up with nothing except for "making a flowchart that describes how it should function", however, isn't that how all software development should be done? So my question is: "What is a hand written lexer?"

    Read the article

  • SQL Server v.Next (Denali) : Another SSMS bug that should be fixed

    - by AaronBertrand
    Sorry to call this out in a separate post (I talked about a bunch of SSMS Connect items the other day), but Aaron Nelson ( blog | twitter ) jogged my memory today about an issue that has gone unfixed for years: the custom coloring for Registered Servers is neither consistent nor global. For one of my servers, I've chosen a red color to show in the status bar. Let's pretend this is a production server, and I want the red to remind me to use caution. I can set this up by right-clicking a Registered...(read more)

    Read the article

  • acer 1551 suspend/hibernate bug?

    - by user11050
    I put 64-bit Ubuntu 10.10 on my Acer 1551. It does not want to wake from suspend or hibernate. I have tried several different solutions to the apparent problem, but none have worked for me. Similar problem: Acer Aspire 1551 Will Not Wake After Suspend I tried what it suggested, no go. Also, I think it is taking my acer a while to boot-up what can I look into that may be causing it? I am new to ubuntu, and would like some help! thanks

    Read the article

  • Cardinality Estimation Bug with Lookups in SQL Server 2008 onward

    - by Paul White
    Cost-based optimization stands or falls on the quality of cardinality estimates (expected row counts).  If the optimizer has incorrect information to start with, it is quite unlikely to produce good quality execution plans except by chance.  There are many ways we can provide good starting information to the optimizer, and even more ways for cardinality estimation to go wrong.  Good database people know this, and work hard to write optimizer-friendly queries with a schema and metadata (e.g. statistics) that reduce the chances of poor cardinality estimation producing a sub-optimal plan.  Today, I am going to look at a case where poor cardinality estimation is Microsoft’s fault, and not yours. SQL Server 2005 SELECT th.ProductID, th.TransactionID, th.TransactionDate FROM Production.TransactionHistory AS th WHERE th.ProductID = 1 AND th.TransactionDate BETWEEN '20030901' AND '20031231'; The query plan on SQL Server 2005 is as follows (if you are using a more recent version of AdventureWorks, you will need to change the year on the date range from 2003 to 2007): There is an Index Seek on ProductID = 1, followed by a Key Lookup to find the Transaction Date for each row, and finally a Filter to restrict the results to only those rows where Transaction Date falls in the range specified.  The cardinality estimate of 45 rows at the Index Seek is exactly correct.  The table is not very large, there are up-to-date statistics associated with the index, so this is as expected. The estimate for the Key Lookup is also exactly right.  Each lookup into the Clustered Index to find the Transaction Date is guaranteed to return exactly one row.  The plan shows that the Key Lookup is expected to be executed 45 times.  The estimate for the Inner Join output is also correct – 45 rows from the seek joining to one row each time, gives 45 rows as output. The Filter estimate is also very good: the optimizer estimates 16.9951 rows will match the specified range of transaction dates.  Eleven rows are produced by this query, but that small difference is quite normal and certainly nothing to worry about here.  All good so far. SQL Server 2008 onward The same query executed against an identical copy of AdventureWorks on SQL Server 2008 produces a different execution plan: The optimizer has pushed the Filter conditions seen in the 2005 plan down to the Key Lookup.  This is a good optimization – it makes sense to filter rows out as early as possible.  Unfortunately, it has made a bit of a mess of the cardinality estimates. The post-Filter estimate of 16.9951 rows seen in the 2005 plan has moved with the predicate on Transaction Date.  Instead of estimating one row, the plan now suggests that 16.9951 rows will be produced by each clustered index lookup – clearly not right!  This misinformation also confuses SQL Sentry Plan Explorer: Plan Explorer shows 765 rows expected from the Key Lookup (it multiplies a rounded estimate of 17 rows by 45 expected executions to give 765 rows total). Workarounds One workaround is to provide a covering non-clustered index (avoiding the lookup avoids the problem of course): CREATE INDEX nc1 ON Production.TransactionHistory (ProductID) INCLUDE (TransactionDate); With the Transaction Date filter applied as a residual predicate in the same operator as the seek, the estimate is again as expected: We could also force the use of the ultimate covering index (the clustered one): SELECT th.ProductID, th.TransactionID, th.TransactionDate FROM Production.TransactionHistory AS th WITH (INDEX(1)) WHERE th.ProductID = 1 AND th.TransactionDate BETWEEN '20030901' AND '20031231'; Summary Providing a covering non-clustered index for all possible queries is not always practical, and scanning the clustered index will rarely be optimal.  Nevertheless, these are the best workarounds we have today. In the meantime, watch out for poor cardinality estimates when a predicate is applied as part of a lookup. The worst thing is that the estimate after the lookup join in the 2008+ plans is wrong.  It’s not hopelessly wrong in this particular case (45 versus 16.9951 is not the end of the world) but it easily can be much worse, and there’s not much you can do about it.  Any decisions made by the optimizer after such a lookup could be based on very wrong information – which can only be bad news. If you think this situation should be improved, please vote for this Connect item. © 2012 Paul White – All Rights Reserved twitter: @SQL_Kiwi email: [email protected]

    Read the article

< Previous Page | 13 14 15 16 17 18 19 20 21 22 23 24  | Next Page >