Search Results

Search found 87901 results on 3517 pages for 'exchange server'.

facebook twitter digg google delicious technorati stumbleupon myspace wordpress linkedin gmail igoogle windows live tumblr viadeo yahoo buzz yahoo mail yahoo bookmarks favorites email print

Page 367/3517 | < Previous Page | 363 364 365 366 367 368 369 370 371 372 373 374  | Next Page >

  • Child sProc cannot reference a Local temp table created in parent sProc

    - by John Galt
    On our production SQL2000 instance, we have a database with hundreds of stored procedures, many of which use a technique of creating a #TEMP table "early" on in the code and then various inner stored procedures get EXECUTEd by this parent sProc. In SQL2000, the inner or "child" sProc have no problem INSERTing into #TEMP or SELECTing data from #TEMP. In short, I assume they can all refer to this #TEMP because they use the same connection. In testing with SQL2008, I find 2 manifestations of different behavior. First, at design time, the new "intellisense" feature is complaining in Management Studio EDIT of the child sProc that #TEMP is an "invalid object name". But worse is that at execution time, the invoked parent sProc fails inside the nested child sProc. Someone suggested that the solution is to change to ##TEMP which is apparently a global temporary table which can be referenced from different connections. That seems too drastic a proposal both from the amount of work to chase down all the problem spots as well as possible/probable nasty effects when these sProcs are invoked from web applications (i.e. multiuser issues). Is this indeed a change in behavior in SQL2005 or SQL2008 regarding #TEMP (local temp tables)? We skipped 2005 but I'd like to learn more precisely why this is occuring before I go off and try to hack out the needed fixes. Thanks.

    Read the article

  • MDX Filter a dimention with another dimension

    - by John Hartsock
    To all excuse my ignorance, but im new to MDX below is the query im trying to run SELECT NON EMPTY { [Measures].[Task Burn Down] } ON COLUMNS, NON EMPTY { ([Calendar].[Date].[Date].ALLMEMBERS * [Priority].[Priority Code].[Priority Code].ALLMEMBERS ) } DIMENSION PROPERTIES MEMBER_CAPTION, MEMBER_UNIQUE_NAME ON ROWS FROM ( SELECT ( { Filter([Calendar].[Date].AllMembers, [Calendar].[Date].CurrentMember <= [Sprint].[Sprint].CurrentMember.Properties("Stop Date")) } ) ON COLUMNS FROM ( SELECT ( { [Sprint].[Sprint].AllMembers } ) ON COLUMNS FROM ( SELECT ( { [Team].[Team].AllMembers } ) ON COLUMNS FROM [Backlog2] ) ) ) WHERE ( [Team].[Team].&[27], [Sprint].[Sprint].&[9] ) CELL PROPERTIES VALUE, BACK_COLOR, FORE_COLOR, FORMATTED_VALUE, FORMAT_STRING, FONT_NAME, FONT_SIZE, FONT_FLAGS When I run this I expect to get 1 value for [Measures].[Task Burn Down] for each day of the month. But I only want one month. My Sprint dimension has a stop date and I do not want to return any values after the stop date. What am I doing wrong?

    Read the article

  • Dynamically call a stored procedure from another stored procedure

    - by Greg
    I want to be able to pass in the name of a stored procedure as a string into another stored procedure and have it called with dynamic parameters. I'm getting an error though. Specifically I've tried: create procedure test @var1 varchar(255), @var2 varchar(255) as select 1 create procedure call_it @proc_name varchar(255) as declare @sp_str varchar(255) set @sp_str = @proc_name + ' ''a'',''b''' print @sp_str exec @sp_str exec call_it 'test' So procedure call_it should call procedure test with arguments 'a', and 'b'. When I run the above code I get: Msg 2812, Level 16, State 62, Procedure call_it, Line 6 Could not find stored procedure 'test 'a','b''. However, running test 'a','b' works fine.

    Read the article

  • Repeatedly execute a stored procedure

    - by manivineet
    I have a situation where I need to repeatedly execute a stored procedure Now this procedure (spMAIN) has a cursor inside which looks for a value from a table as T1,which has the following structure ID Status ---- -------- 1 New 2 New 3 success 4 Error now the cursor looks for all rows with a status of 'New' Now while processing , if that instance of the cursor encounters an error, another SP say spError needs to be called, the 'Status' column in T1 needs to be updated to 'Error' and spMAIN needs to be called again which again repeats the process, looking for rows with 'new' how do I do it? Also, also, while we are at it, what if an SP has other SPs inside it and if any of those SP raises an error, same thing needs to be done, the T1 table needs to be updated ('Error') and spMAIN needs to be called again. can you also recommend something ? here's some code ALTER PROC zzSpMain AS BEGIN DECLARE @id INT BEGIN TRY IF EXISTS ( SELECT * FROM dbo.zzTest WHERE istatus = 'new' ) BEGIN DECLARE c CURSOR FOR SELECT id FROM zztest WHERE istatus = 'new' OPEN c FETCH NEXT FROM c INTO @id WHILE @@FETCH_STATUS = 0 BEGIN PRINT @id IF @id = 2 BEGIN UPDATE zztest SET istatus = 'error' WHERE id = @id RAISERROR ( 'Error occured', 16, 1 ) END UPDATE zztest SET istatus = 'processed' WHERE id = @id FETCH NEXT FROM c INTO @id END CLOSE c DEALLOCATE c END END TRY begin CATCH EXEC zzSpError END CATCH END

    Read the article

  • SQL2k8 T-SQL: Output into XML file

    - by Nai
    I have two tables Table Name: Graph UID1 UID2 ----------- 12 23 12 32 41 51 32 41 Table Name: Profiles NodeID UID Name ----------------- 1 12 Robs 2 23 Jones 3 32 Lim 4 41 Teo 5 51 Zacks I want to get an xml file like this: <graph directed="0"> <node id="1"> <att name="UID" value="12"/> <att name="Name" value="Robs"/> </node> <node id="2"> <att name="UID" value="23"/> <att name="Name" value="Jones"/> </node> <node id="3"> <att name="UID" value="32"/> <att name="Name" value="Lim"/> </node> <node id="4"> <att name="UID" value="41"/> <att name="Name" value="Teo"/> </node> <node id="5"> <att name="UID" value="51"/> <att name="Name" value="Zacks"/> </node> <edge source="12" target="23" /> <edge source="12" target="32" /> <edge source="41" target="51" /> <edge source="32" target="41" /> </graph> Thanks very much!

    Read the article

  • SQL Server 2005 - query with case statement

    - by user329266
    Trying to put a single query together to be used eventually in a SQL Server 2005 report. I need to: Pull in all distinct records for values in the "eventid" column for a time frame - this seems to work. For each eventid referenced above, I need to search for all instances of the same eventid to see if there is another record with TaskName like 'review1%'. Again, this seems to work. This is where things get complicated: For each record where TaskName is like review1, I need to see if another record exists with the same eventid and where TaskName='End'. Utimately, I need a count of how many records have TaskName like 'review1%', and then how many have TaskName like 'review1%' AND TaskName='End'. I would think this could be accomplished by setting a new value for each record, and for the eventid, if a record exists with TaskName='End', set to 1, and if not, set to 0. The query below seems to accomplish item #1 above: SELECT eventid, TimeStamp, TaskName, filepath FROM (SELECT eventid, TimeStamp, filepath, TaskName, ROW_NUMBER() OVER(PARTITION BY eventid ORDER BY TimeStamp DESC) AS seq FROM eventrecords where ((TimeStamp >= '2010-4-1 00:00:00.000') and (TimeStamp <= '2010-4-21 00:00:00.000'))) AS T WHERE seq = 1 order by eventid And the query below seems to accomplish #2: SELECT eventid, TimeStamp, TaskName, filepath FROM (SELECT eventid, TimeStamp, filepath, TaskName, ROW_NUMBER() OVER(PARTITION BY eventid ORDER BY TimeStamp DESC) AS seq FROM eventrecords where ((TimeStamp >= '2010-4-1 00:00:00.000') and (TimeStamp <= '2010-4-21 00:00:00.000')) and TaskName like 'Review1%') AS T WHERE seq = 1 order by eventid This will bring back the eventid's that also have a TaskName='End': SELECT eventid, TimeStamp, TaskName, filepath FROM (SELECT eventid, TimeStamp, filepath, TaskName, ROW_NUMBER() OVER(PARTITION BY eventid ORDER BY TimeStamp DESC) AS seq FROM eventrecords where ((TimeStamp >= '2010-4-1 00:00:00.000') and (TimeStamp <= '2010-4-21 00:00:00.000')) and TaskName like 'Review1%') AS T WHERE seq = 1 and eventid in (Select eventid from eventrecords where TaskName = 'End') order by eventid So I've tried the following to TRY to accomplish #3: SELECT eventid, TimeStamp, TaskName, filepath FROM (SELECT eventid, TimeStamp, filepath, TaskName, ROW_NUMBER() OVER(PARTITION BY eventid ORDER BY TimeStamp DESC) AS seq FROM eventrecords where ((TimeStamp >= '2010-4-1 00:00:00.000') and (TimeStamp <= '2010-4-21 00:00:00.000')) and TaskName like 'Review1%') AS T WHERE seq = 1 and case when (eventid in (Select eventid from eventrecords where TaskName = 'End') then 1 else 0) as bit end order by eventid When I try to run this, I get: "Incorrect syntax near the keyword 'then'." Not sure what I'm doing wrong. Haven't seen any examples anywhere quite like this. I should mention that eventrecords has a primary key, but it doesn't seem to help anything when I include it, and I am not permitted to change the table. (ugh) I've received one suggestion to use a cursor and temporary table, but am not sure how badley that would bog down performance when the report is running. Thanks in advance.

    Read the article

  • How to call an extended procedure from a function

    - by Alan Featherston Lago
    hi im having trouble trying to get the following function to work. CREATE FUNCTION test ( @nt_group VARCHAR(128) ) RETURNS @nt_usr TABLE ( [name] [nchar](128) NULL , [type] [char](8) NULL , [privilege] [char](9) NULL , [mapped login name] [nchar](128) NULL , [permission path] [nchar](128) NULL ) AS BEGIN INSERT INTO @nt_usr EXEC master.dbo.xp_logininfo 'DOMAIN\USER', @nt_group RETURN END As far as i know i should be allowed to call an extended stored procedure, im getting the following error Mes 443, Level 16, State 14 Could it be that xp_logininfo might return different result sets depending on the parameters? When i use openquery i can overcome this by setting this: SET FMTONLY OFF. Does anyone know if there's a similar workaround for my problem?

    Read the article

  • How to import data in SQL Compact Edition?

    - by Peter
    I don't seem to find a tool for it, nor an odbc driver. Thanks UPDATE : I'm aware of the sql scripting possibilities. But than again : how to script a sql 2k table? (not just ddl, but data also?) Of course you can write this all by yourself, but importing data into CE cannot be such a hassle, or can it ? UPDATE2 : I don't seem to be able to choose the right dialect for inserting

    Read the article

  • SQL to get list of dates as well as days before and after without duplicates

    - by Nathan Koop
    I need to display a list of dates, which I have in a table SELECT mydate AS MyDate, 1 AS DateType FROM myTable WHERE myTable.fkId = @MyFkId; Jan 1, 2010 - 1 Jan 2, 2010 - 1 Jan 10, 2010 - 1 No problem. However, I now need to display the date before and the date after as well with a different DateType. Dec 31, 2009 - 2 Jan 1, 2010 - 1 Jan 2, 2010 - 1 Jan 3, 2010 - 2 Jan 9, 2010 - 2 Jan 10, 2010 - 1 Jan 11, 2010 - 2 I thought I could use a union SELECT MyDate, DateType FROM ( SELECT mydate - 1 AS MyDate, 2 AS DateType FROM myTable WHERE myTable.fkId = @MyFkId; UNION SELECT mydate + 1 AS MyDate, 2 AS DateType FROM myTable WHERE myTable.fkId = @MyFkId; UNION SELECT mydate AS MyDate, 1 AS DateType FROM myTable WHERE myTable.fkId = @MyFkId; ) AS myCombinedDateTable This however includes duplicates of the original dates. Dec 31, 2009 - 2 Jan 1, 2009 - 2 Jan 1, 2010 - 1 Jan 2, 2010 - 2 Jan 2, 2010 - 1 Jan 3, 2010 - 2 Jan 9, 2010 - 2 Jan 10, 2010 - 1 Jan 11, 2010 - 2 How can I best remove these duplicates? I am considering a temporary table, but am unsure if that is the best way to do it. This also appears to me that it may provide performance issues as I am running the same query three separate times. What would be the best way to handle this request?

    Read the article

  • switch case in where clause

    - by Nimesh
    hi, i need to check three conditions: if @filter = 1 { **select * from employeestable where rating is not null** } else if @filter = 2 { **select * from employeestable where rating is null** } else { **select * from employeestable** } This i need to do using a case statement. now i have more than 30 lines of query, if i use case i can reduce my code upto 70% Please let mek now how can i do this.

    Read the article

  • Cross Join 'n' times a table

    - by SDReyes
    It is possible to write a generic function/procedure/select/somethingElse to cross-join a table against himself 'n' times? (yes, 'n' is a given parameter : ) How would you do it? Example Having this table: Value ------- 1 2 3 cross join it 2 times, would return: Value | Value ------------------ 1 1 1 2 1 3 2 1 2 2 2 3 3 1 3 2 3 3

    Read the article

  • Full text search on multiple fields in different tables

    - by crisgomez
    Hi, I want to create a a full text search usng ms sql 2005. here is the structure of my table User-contains fields of Id, fname,lname,email,alternativeemail Attachment-contains fields of id, name,category,filenameorurl Certification-contains of fileds of id, title,school,sdate,edate EducationalBackground-contain of fields of, id,qualification,sdate,edate,school,fieldofstudy EmploymentDetails-contain of fields position, positionlevelid,specializationid,description now the relationship of user to the other table is one to many relationship. now how can I create an effective query to have quick full text search and return the values of the different tables?any help please

    Read the article

  • add ANOTHER primary key to a table which is UNIQUE

    - by gdubs
    so im having problems with adding another primary key to my table. i have 3 columns: 1. Account ID (Identity) 2. EmailID 3. Data field when i made the table i had this to make the Account ID and the Email ID unique PRIMARY KEY (AccountID, EmailID) i thought that would make my emailid unique, but then after i tried inserting another row with the same emailid it went through. so i thought i missed something out. now for my question: IF, i had to use alter, How do i alter the table/PK Constraint to modify the EmailID field and make it Unique IF i decided to drop the table and made a new one, how do i make those two primary keys uniqe? Thanks a bunch!!

    Read the article

  • FTS: Searching across multiple fields 'intelligently'

    - by Wild Thing
    Hi, I have a SP using FTS (Full Text Search). I want searches across multiple fields, 'intelligently' ranking results based on the weights I assign. Consider a search on a view fetching data from tables: Book, Author and Genre. Now, I want the searcher to be able to do: "Ludlum Fiction", "Robert Ludlum Bourne", "Bourne Ludlum", etc. Unfortunately, the only way I have been able to do that at present is this: http://pastebin.com/fdce11ff This is pretty bad, because I am manually breaking up the search string. I know I am doing this completely the wrong way, but can't figure out the right way to search across multiple fields in FTS. Can somebody help please?

    Read the article

  • filling in the holes in the result of a query

    - by ????? ????????
    my query is returning: +------+------+------+------+------+------+------+-------+------+------+------+------+-----+ | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | Bla | +------+------+------+------+------+------+------+-------+------+------+------+------+-----+ | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 2 | 0 | 13 | | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 2 | 0 | 0 | 14 | | 0 | 0 | 0 | 0 | 0 | 9 | 0 | 0 | 0 | 0 | 8 | 37 | 29 | | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 374 | 30 | | 0 | 0 | 1 | 0 | 78 | 2 | 4 | 8 | 57 | 169 | 116 | 602 | 31 | | 156 | 255 | 79 | 75 | 684 | 325 | 289 | 194 | 407 | 171 | 584 | 443 | 32 | | 1561 | 2852 | 2056 | 796 | 2004 | 1755 | 879 | 1052 | 1490 | 1683 | 2532 | 2381 | 33 | | 4167 | 3841 | 4798 | 3399 | 4132 | 5849 | 3157 | 4381 | 4424 | 4487 | 4178 | 5343 | 34 | | 5472 | 5939 | 5768 | 4150 | 7483 | 6836 | 6346 | 6288 | 6850 | 7155 | 5706 | 5231 | 35 | | 5749 | 4741 | 5264 | 4045 | 6544 | 7405 | 7524 | 6625 | 6344 | 5508 | 6513 | 3854 | 36 | | 5464 | 6323 | 7074 | 4861 | 7244 | 6768 | 6632 | 7389 | 8077 | 8745 | 6738 | 5039 | 37 | | 5731 | 7205 | 7476 | 5734 | 9103 | 9244 | 7339 | 8970 | 9726 | 9089 | 6328 | 5512 | 38 | | 7262 | 6149 | 8231 | 6654 | 9886 | 9834 | 9306 | 10065 | 9983 | 9984 | 6738 | 5806 | 39 | | 5886 | 6934 | 7137 | 6978 | 9034 | 9155 | 7389 | 9437 | 9711 | 8665 | 6593 | 5337 | 40 | +------+------+------+------+------+------+------+-------+------+------+------+------+-----+ as you can see the BLA column starts from 13. i want it to start from 1, then 2, then 3 etc......I do not want any gaps in the data. The reason there are gaps is because all of the months are 0 for that specific bla how do i get the result set to include ALL values for BLA, even ones that will yield 0 for the months? here are the desired results: +-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+ | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | Bla | +-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+ | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 2 | | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 3 | | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 4 | | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 5 | | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 6 | | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 7 | | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 8 | | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 9 | | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 10 | | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 11 | | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 12 | | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 13 | | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 2 | 0 | 0 | 14 | | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 15 | | … | … | … | … | … | … | … | … | … | … | … | … | … | +-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+ here's my query: WITH CTE AS ( select sum(case when datepart(month,[datetime entered]) = 1 then 1 end) as Jan, sum(case when datepart(month,[datetime entered]) = 2 then 1 end) as Feb, sum(case when datepart(month,[datetime entered]) = 3 then 1 end) as Mar, sum(case when datepart(month,[datetime entered]) = 4 then 1 end) as Apr, sum(case when datepart(month,[datetime entered]) = 5 then 1 end) as May, sum(case when datepart(month,[datetime entered]) = 6 then 1 end) as Jun, sum(case when datepart(month,[datetime entered]) = 7 then 1 end) as Jul, sum(case when datepart(month,[datetime entered]) = 8 then 1 end) as Aug, sum(case when datepart(month,[datetime entered]) = 9 then 1 end) as Sep, sum(case when datepart(month,[datetime entered]) = 10 then 1 end) as Oct, sum(case when datepart(month,[datetime entered]) = 11 then 1 end) as Nov, sum(case when datepart(month,[datetime entered]) = 12 then 1 end) as Dec, DATEPART(yyyy,[datetime entered]) as [Year], bla= CASE WHEN datediff(d, CAST([datetime entered] as DATE), CAST([datetime completed] as DATE))*24 + CONVERT(CHAR(2),[datetime completed],108) >191 THEN 192 ELSE datediff(d, CAST([datetime entered] as DATE), CAST([datetime completed] as DATE))*24 + CONVERT(CHAR(2),[datetime completed],108) END --,datediff(d, CAST([datetime entered] as DATE), CAST([datetime completed] as DATE)) AS Sort_Days, --DATEPART(hour, [datetime completed] ) AS Sort_Hours from [TurnAround] group by datediff(d, CAST([datetime entered] as DATE), CAST([datetime completed] as DATE))*24 + CONVERT(CHAR(2),[datetime completed],108), DATEPART(yyyy,[datetime entered]) , [datetime entered] --[DateTime Completed] ) SELECT ISNULL(SUM(Jan),0) Jan, ISNULL(SUM(Feb),0) Feb, ISNULL(SUM(Mar),0) Mar, ISNULL(SUM(Apr),0) Apr, ISNULL(SUM(May),0) May, ISNULL(SUM(Jun),0) Jun, ISNULL(SUM(Jul),0) Jul, ISNULL(SUM(Aug),0) Aug, ISNULL(SUM(Sep),0) Sep, ISNULL(SUM(Oct),0) Oct, ISNULL(SUM(Nov),0) Nov, ISNULL(SUM(Dec),0) Dec, [year], --,Sort_Hours, --Sort_Days, A.RN Bla FROM ( SELECT *, RN=ROW_NUMBER() OVER(ORDER BY object_id) FROM sys.all_objects) A LEFT JOIN CTE B ON A.RN = CASE WHEN B.Bla > 191 THEN 192 ELSE B.Bla END WHERE A.RN BETWEEN 1 AND 192 GROUP BY A.RN,[year]

    Read the article

  • SQL Express Edition, SQL Compact Editin and SQLCMD for learning purpose

    - by Mil
    Hi, I want to learn programming in SQL from some SQL tutorial sites of which I heard of here but I need some environment for executing query's. I think I have both SQL CE and SQL EE installed on my computer but I have some doubts about these DBMS and I don't know exactly how to use SQLCMD utility so I hope someone here will have time and will to explain me the following: Since running sqlcmd -S.\sqlexpress at command prompt command gives "1" prompt I assume I have SQL express installed but anyway how can I be sure what I have installed on my machine since I cannot find in installed programs SQL Express Edition name? Can I ship and use database with my C# (VC# Express) application which was created with SQL EE (embedded?)? How can use sqlcmd for learning SQL, that is by issuing commands like create, use, select..., again emphasize is on learning SQL I do not want to run scripts but use interactive command prompt like with MySQL (since I want to use SQL I would pretty much like to avoid graphical tools for DBMS)? Please tell me if you have some other advice regarding as to what should I better use in learning how to program in SQL or should I stick with the above for now. Thanks in advance.

    Read the article

  • Problem with Full text Searching

    - by devendra
    I am searching in resumes weather the word is exist or not i am using the below query Case1: select top(10) c_resume_text from sntbl_candidates where contains(c_resume_text,'"a/dm"') in the above example only it is not working properly .It showing resumes even though there is no text like that. In Messages i am getting the following message. Informational: The full-text search condition contained noise word(s). if i try with Case 2: select top(10) c_resume_text from sntbl_candidates where contains(c_resume_text,'"a/d') i am getting proper results in case 2 can any one suggest me what to do. Thanks

    Read the article

  • Is READ UNCOMMITTED / NOLOCK safe in this situation?

    - by Ben Challenor
    I know that snapshot isolation would fix this problem, but I'm wondering if NOLOCK is safe in this specific case so that I can avoid the overhead. I have a table that looks something like this: drop table Data create table Data ( Id BIGINT NOT NULL, Date BIGINT NOT NULL, Value BIGINT, constraint Cx primary key (Date, Id) ) create nonclustered index Ix on Data (Id, Date) There are no updates to the table, ever. Deletes can occur but they should never contend with the SELECT because they affect the other, older end of the table. Inserts are regular and page splits to the (Id, Date) index are extremely common. I have a deadlock situation between a standard INSERT and a SELECT that looks like this: select top 1 Date, Value from Data where Id = @p0 order by Date desc because the INSERT acquires a lock on Cx (Date, Id; Value) and then Ix (Id, Date), but the SELECT acquires a lock on Ix (Id, Date) and then Cx (Date, Id; Value). This is because the SELECT first seeks on Ix and then joins to a seek on Cx. Swapping the clustered and non-clustered index would break this cycle, but it is not an acceptable solution because it would introduce cycles with other (more complex) SELECTs. If I add NOLOCK to the SELECT, can it go wrong in this case? Can it return: More than one row, even though I asked for TOP 1? No rows, even though one exists and has been committed? Worst of all, a row that doesn't satisfy the WHERE clause? I've done a lot of reading about this online, but the only reproductions of over- or under-count anomalies I've seen (one, two) involve a scan. This involves only seeks. Jeff Atwood has a post about using NOLOCK that generated a good discussion. I was particularly interested in a comment by Rick Townsend: Secondly, if you read dirty data, the risk you run is of reading the entirely wrong row. For example, if your select reads an index to find your row, then the update changes the location of the rows (e.g.: due to a page split or an update to the clustered index), when your select goes to read the actual data row, it's either no longer there, or a different row altogether! Is this possible with inserts only, and no updates? If so, then I guess even my seeks on an insert-only table could be dangerous. Update: I'm trying to figure out how snapshot isolation works. It seems to be row-based, where transactions read the table (with no shared lock!), find the row they are interested in, and then see if they need to get an old version of the row from the version store in tempdb. But in my case, no row will have more than one version, so the version store seems rather pointless. And if the row was found with no shared lock, how is it different to just using NOLOCK?

    Read the article

facebook twitter digg google delicious technorati stumbleupon myspace wordpress linkedin gmail igoogle windows live tumblr viadeo yahoo buzz yahoo mail yahoo bookmarks favorites email print

< Previous Page | 363 364 365 366 367 368 369 370 371 372 373 374  | Next Page >