I'm trying to create a SQL server 2000 function that returns a scalar value, but I keep getting the error "Return statements in scalar valued functions must include an argument". Online clarification of this error message is no help at all.I've tried all sorts of combinations of the following, without much luck. Can someone point out my dim-witted mistake, please?ALTER FUNCTION dbo.intCoursesPublic (@intCatID as int) RETURNS intASBEGIN RETURN SELECT COUNT(intCourseID) AS Expr1 FROM dbo.tbl_guru_course_list WHERE (intCatID = @intCatID)END
So I was creating a new table-valued function today which queries some data from a preexisting table. Since this is my first table-valued function, I decided to check out some of the examples and see what I can figure out.
One particular example helped me out a bit until I ran into some data access issues... http://msdn2.microsoft.com/en-us/library/ms165054.aspx
So I create my function:
[SqlFunction(DataAccess = DataAccessKind.Read,SystemDataAccess=SystemDataAccessKind.Read,FillRowMethodName = "FillMyRow",TableDefinition ="p1 int, p2 int"] public static IEnumerable getMyTable() { using (SqlConnection conn = ....) { using (SqlCommand command = conn.CreateCommand()) { ///.... populate command text, open connection using (SqlDataReader rdr = command.ExecuteReader()) { while (rdr.Read()) { customObject1 o = new customObject1(); ///... populate o's parameters from reader ... yield return o; } } } }
public static void FillMyRow( object source, out int p1, out int p2) { customObject1 f = (customObject1)source; p1 = f.p1; p2 = f.p2; }
Notice, this example yield returns the value o upon each iteration of the reader. Despite the fact that the DataAccess is set to Read I still get the error...
An error occurred while getting new row from user defined Table Valued Function :
System.InvalidOperationException: Data access is not allowed in this context. Either the context is a function or method not marked with DataAccessKind.Read or SystemDataAccessKind.Read, is a callback to obtain data from FillRow method of a Table Valued Function, or is a UDT validation method.
I did however get past this error, by creating a collection of customObject1, populated it within the while(rdr.Read()) loop, then return the collection after closing the connection, command and reader.
I assume this error has something to do with the fact that you can't yield return results from within an open reader. Is this error right though in this case? Whats causing it to throw a InvOp Exception? Or is this a bug?
From the SQL Server documentation : "The input parameters and the type returned from a SVF can be any of the scalar data types supported by SQL Server, except rowversion, text, ntext, image, timestamp, table, or cursor"This is a problem for me. Here's what I'm trying to do :I have an NTEXT field in one of my tables. I want to run regular expressions on this field, and return the results from a stored procedure. Since SQL Server doesn't provide facilities to perform regular expressions, I need to use an SQLCLR function. I would have no problem doing this if my field was nvarchar. However, this field needs to be variable in length - I cannot set an upper bound. This is why I'm using NTEXT and not nvarchar in the first place.Is there a solution to this problem? I can't imagine that I'm the only person who wants to pass strings of arbitrary size to an SQLCLR function.
I have an assembly that contains the following function:
Public Class Lookup
<SqlFunction()> _
Public Shared Function MyTest() As Integer
Return System.Data.SqlTypes.SqlInt64.Null
End Function
End Class
Then in SSMS:
CREATE ASSEMBLY IRS_MyTest
FROM '\machine empmyAssembly.dll'
GO
CREATE FUNCTION dbo.MyTest() RETURNS INT
AS EXTERNAL NAME IRS_MyTest.[MyClass].MyTest
GO
when I run:
SELECT dbo.MyTest()
the following is returned:
Msg 6522, Level 16, State 2, Line 1
A .NET Framework error occurred during execution of user defined routine or aggregate 'MyTest':
System.Data.SqlTypes.SqlNullValueException: Data is Null. This method or property cannot be called on Null values.
System.Data.SqlTypes.SqlNullValueException:
at System.Data.SqlTypes.SqlInt64.get_Value()
at System.Data.SqlTypes.SqlInt64.op_Explicit(SqlInt64 x)
at Informed.DCLG.IRS.SQL.IncidentTransform.Lookup.MyTest()
Can anyone please advise on how to return back a null value. Currently, my only other option is to return nothing (actually returns 0) and then wrap this up to convert the value to null - not ideal.
public static void FillRows(Object obj, out SqlString SOCIETA, out SqlString CLIENTE, out SqlString NUMEROCONTRATTO, out SqlDateTime FIRMA, out SqlDateTime CHIUSURA, SqlDouble AUTORIZZATO)
Whe I try to deploy my function, I get the following error:
Error 1 Function signature of "FillRow" method (as designated by SqlFunctionAttribute.FillRowMethodName) does not match SQL declaration for table valued CLR function 'dbf_Create_RiepilogoAccordi' due to column 6. CM.Reports.SIA.RiepilogoAccordi
I get this error whichever combination of name/value I use for column 6
Ok, I'm pretty knowledgable about T-SQL, but I've hit something that seems should work, but just doesn't... I'm writing a stored procedure that needs to use the primary key fields of a table that is being passed to me so that I can generate what will most likely be a dynamically generated SQL statement and then execute it. So the first thing I do, is I need to grab the primary key fields of the table. I'd rather not go down to the base system tables since we may (hopefully) upgrade this one SQL 2000 machine to 2005 fairly soon, so I poke around, and find sp_pkeys in the master table. Great. I pass in the table name, and sure enough, it comes back with a record set, 1 row per column. That's exactly what I need. Umm... This is the part where I'm at a loss. The stored procedure outputs the resultset as a resultset (Not as an output param). Now I want to use that list in my stored procedure, thinking that if the base tables change, Microsoft will change the stored procedure accordingly, so even after a version upgrade my stuff SHOULD still work. But... How do I use the resultset from the stored procedure? You can't reference it like a table-valued function, nor can you 'capture' the resultset for use using the syntax like: DECLARE @table table@table=EXEC sp_pkeys MyTable That of course just returns you the RETURN_VALUE instead of the resultset it output. Ugh. Ok, so I finally decide to just bite the bullet, and I grab the code from sp_pkeys and make my own little function called fn_pkeys. Since I might also want to be able to 'force' the primary keys (Maybe the table doesn't really have one, but logically it does), I decide it'll pass back a comma-delimited varchar of columns that make up the primary key. Ok, I test it and it works great. Now, I'm happily going along and building my routine, and realize, hey, I don't really want that in a comma-delimited varchar, I want to use it in one of my queries, and I have this nice little table-valued function I call split, that takes a comma-delimited varchar, and returns a table... So I preceed to try it out... SELECT *FROM Split(fn_pkeys('MyTable'),DEFAULT) Syntax Error. Ugh. Eventually, I even try: SELECT *FROM Split(substring('abc,def',2,6),DEFAULT) Syntax Error. Hmm...What am I doing wrong here, or can't you use a scalar-valued function as a parameter into a table-valued function? SELECT *FROM Split('bc,def',DEFAULT) works just fine. So my questions are: Is there any way to programmatically capture a resultset that is being output from a stored procedure for use in the stored procedure that called it? Is there any way to pass a scalar-valued function as a parameter into a table-valued function? Oh, this works as well as a work around, but I'm more interested in if there is a way without having to workaround: DECLARE @tmp varchar(8000) SET @tmp=(SELECT dbo.fn_pkeys('MyTable')) SELECT * FROM Split(@tmp,DEFAULT)
Select statements included within a function cannot return data to a client.
Is this a proper way to include a CTE in a function?
USE [DB1] GO SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO CREATE FUNCTION [dbo].[_Pink_FN_StartingDatePLGeographical](@StartingDate AS DATETIME) RETURNS NVARCHAR(20)
Hi, I'm having trouble with this multi-statement table-valued function:
ALTER FUNCTION MakeArDetail ( -- Add the parameters for the function here @dateStart DATETIME, @dateEnd DATETIME ) RETURNS @arDetail TABLE ( Insurer VARCHAR(50), NABP INT DEFAULT 0, Claim MONEY DEFAULT 0, Payment MONEY DEFAULT 0, NumRx CHAR(7), PatientName VARCHAR(50), Paid030 MONEY DEFAULT 0, Paid3160 MONEY DEFAULT 0, Paid6190 MONEY DEFAULT 0, Paid91120 MONEY DEFAULT 0, Paid121 MONEY DEFAULT 0 ) AS BEGIN DECLARE @arTemp TABLE ( Insurer VARCHAR(50), NABP INT DEFAULT 0, Claim MONEY DEFAULT 0, Payment MONEY DEFAULT 0, NumRx CHAR(7), PatientName VARCHAR(50), Paid030 MONEY DEFAULT 0, Paid3160 MONEY DEFAULT 0, Paid6190 MONEY DEFAULT 0, Paid91120 MONEY DEFAULT 0, Paid121 MONEY DEFAULT 0 )
INSERT INTO @arTemp SELECT DISTINCT Insurer,NABP,0,0,NumRx,Patient,0,0,0,0,0 FROM Pims; UPDATE @arTemp SET Claim = (SELECT SUM(Pims.AmtReq) FROM Pims WHERE Pims.Insurer = @arTemp.Insurer AND Pims.NABP = @arTemp.NABP AND Pims.NumRx = @arTemp.NumRx );
INSERT INTO @arDetail SELECT * FROM @arTemp RETURN END GO
I get Msg 137, Level 15, State 2, Procedure MakeArDetail, Line 43 Must declare the scalar variable "@arTemp".
I don't understand why SQL thinks @arTemp is a scalar variable which has to be declared. If I don't include the UPDATE command the thing works.
I have a procedure that calls a SVF to convert an xmldocument. The ultimate purpose is to update the xml in a column in a multi-million row table. The xml is stored as varchar(MAX), it was supposed to carry any type of text, initially at least.
My question is: why is the xml-parsing performed inside the function much slower when i pass the xmldocument as type xml than when it is passed as varchar(MAX) and the CAST to xml is within the function? Does processing the xml input parameter in SlowFunction involve expensive crossing of some context border?
The two versions of the SVF (they return the rowcount in this simplified example):
CREATE FUNCTION [dbo].[FastFunction] ( @inDetaljerText varchar(MAX) ) RETURNS int
[Code] ....
The two versions of the SP
CREATE PROCEDURE [dbo].[FastProcedure] AS BEGIN SET NOCOUNT ON; select dbo.FastFunction(al.Detaljer)
I need to call a function to calculate a value. This function accepts a varchar parameter and returns a boolean value. I need to call this function for each row in the dataflow task. I thought I would use an oledb command transformation and for some reason if I say..
'select functioname(?)' as the sqlcommand, it gives me an error message at the design time. In the input/output properties, I have mapped Param_0(external column) to an input column.
I get this erro.."syntax error, ermission violation or other non specific error". Can somebiody please suggest me what's wrong with this and how should I deal this.
I have a view that contains a complex query. A few of the columnscall a function that returns a specific output. I also use a functionto do a join as well.For example:SELECT l.ID, dbo.getStatus(l.ID) AS statusIDFROM tableName ALEFT OUTER JOIN dbo.Status_LKP s ON dbo.getStatus(l.Leg_ID) =s.statusIDFor 800 records, this query takes about 9 seconds. I realize that foreach record, the function is executed on a per row basis, so I amlooking for alternatives.Does anyone know of other ways to accomplish something of the same?Basically I would like to include UDFs in a query and use those UDFsin the where and join clauses.Thanks
I'm studying for the MCDBA test & understand table valued functions but am struggling to find a good use for them... can anyone explain to me why you'd want to use one over a view?
Hi all,I've been using scalar functions as a way to perform some complex datatransformation operations, and I've noticed that scalar functionsreaaaaalllllyyyy sloooowwwwww thiiiiiings dooooooown. I expect slow-down, ofcourse, and would even not be surprised at slow-downs up to a factor of, say50:1, but I'm seeing slow-downs more like 1000:1 or 100000:1. I'm sure itwould actually be faster to actually export a table, use VB to process it,then import it back in.
Hi I am writting Stored Procedures that have to be built on the base of other tables specially created for this purpose, is it better to create these intermediate tables as views or as functions returning tables? I guess views would be lighter on performance as they would not be created on the fly?
Here's my function. The trouble - I can not make ORDER BY the "visits_count", "properties_count", "enquiries_count" fields. May be some one could help me with this?
CREATE FUNCTION [dbo].[GetPagedStatistics] ( @start_index int, @count int, @condition nvarchar(255), @order_field nvarchar(255), @date_from datetime, @date_to datetime ) RETURNS @total_stat TABLE ( username nvarchar(255), first_name nvarchar(255), last_name nvarchar(255), properties_count int, enquiries_count int, visits_count int, id_user int) BEGIN INSERT @total_stat SELECT top (@count) dbo.users.username, dbo.users.first_name, dbo.users.last_name, ISNULL(COUNT(DISTINCT dbo.advertisement.id_advertisement), 0) AS properties_count, ISNULL(COUNT(DISTINCT dbo.enquiry_emails.id_enquiry_email), 0) AS enquiries_count, ISNULL(COUNT(DISTINCT dbo.property_statistics.id_statistics), 0) AS visits_count, dbo.users.id_user FROM dbo.property_statistics RIGHT OUTER JOIN dbo.advertisement RIGHT OUTER JOIN dbo.users ON dbo.advertisement.id_user = dbo.users.id_user LEFT JOIN dbo.enquiry_emails ON dbo.enquiry_emails.id_advertisement = dbo.advertisement.id_advertisement ON dbo.property_statistics.id_advertisement = dbo.advertisement.id_advertisement WHERE 1=@condition and (dbo.advertisement.creation_date <= @date_to and dbo.advertisement.creation_date >= @date_from ) and ( (dbo.enquiry_emails.creation_date <= @date_to and dbo.enquiry_emails.creation_date >= @date_from and dbo.property_statistics.view_date <= @date_to and dbo.property_statistics.view_date >= @date_from ) or (dbo.property_statistics.view_date is null) or (dbo.enquiry_emails.creation_date is null) ) and (ISNULL(dbo.advertisement.id_parent, 0) = 0)
GROUP BY dbo.users.username, dbo.users.first_name, dbo.users.last_name, dbo.users.id_user
order by case when @order_field='username' then dbo.users.username end, case when @order_field='first_name' then dbo.users.first_name end, case when @order_field='last_name' then dbo.users.last_name end, case when @order_field='properties_count' then 1 end, case when @order_field='enquiries_count' then 1 end, case when @order_field='visits_count' then 1 end
Help! Been doing the box step with BOL for several hours , Using tables in Adventureworks to create inline-table-valued function to provide a parameterized view of three JOINS - Have sucessfully created the function but can't figure out where to 'Declare' my variable "@SalesAgentID" need to be able to invoke the function with a particular ID - If you can help me cut this dance short I would REALLY Appreciate it.
Following the horrendous performance of Scalar UDFs demonstrated in this thread URL....(Scalar UDF 10x slower than writing the code, long hand, in the SELECT) I now have the worry that their widespread usage in my code needs to be refactored. I guess pretty much any single-line non-table-valued scalar-function would do
CREATE FUNCTION dbo.FNSafeString ( @strValuevarchar(8000),-- STRING Value to convert to SAFE string @intModeint = 0-- 0=NULL as '[NULL]', 1=NULL as '' ) RETURNS varchar(8000) WITH SCHEMABINDING AS BEGIN RETURN COALESCE(@strValue, CASE WHEN @intMode = 0 THEN '[NULL]' ELSE '' END ) END
Its a little bit too big to really want to have to write it in each place where I need to use it, and its just big enough that it is worth centralising, and it might just be that I could code it more efficiently / different / handle more "edge cases" in the future.I do have even more straightforward UDFs - just to avoid typing something "Lengthy"
CREATE FUNCTION dbo.FNDateMidnight ( @dtDateTimedatetime-- Date/Time Value to adjust ) RETURNS datetime WITH SCHEMABINDING AS BEGIN RETURNDATEADD(Day, DATEDIFF(Day, 0, @dtDateTime)+1, 0) END
Perhaps I should consider changing them to inline table valued functions, or CROSS APPLY them? but I can't think of a way to centralised them and re-use the code WITHOUT a function, can I?
Hi everyone.I'd like to know how stored procedures and table-valued functions compare when it comes to returning a resultant set of data. I know there is a link somewhere but I can't immediately find it.Thanks.
Hello Gurus, I have a stored procedure that gathers data from three tables and joins them, two of the tables need to have different rowcounts set, ie. pull only a certain number of rows from one table and only a certain number of rows from another table... The number of rows it should pull are stored within a table for each. Let me explain.... these tables hold Exchange storage group and mailstore data for a number of servers. Each server has a table entry with the number of child storage groups and each storage group has a table entry with the number of child mailstores. The tables get updated every two minutes via a program. I need to be able to get the most Data with the correct child counts for each server and storage group. I believe that i've found a way to do this with a stored procedure that calls a table-valued function. The table-valued function simply filters down the storage group table to it's number of storage groups, ordered by timestamp. I may be way off here, but i can't tell because both the stored procedure and function check out fine but when i execute the stored procedure it gives me the following error: Cannot find either column "dbo" or the user-defined function or aggregate "dbo.GetExchSGInfo", or the name is ambiguous.
My code is below: Stored Procedure: SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO
Set @SID = (SELECT ServerID FROM dbo.Servers WHERE ServerName = @ServerName) Set @top = (SELECT sum(Children) FROM dbo.ExchangeSG WHERE ServerID = @SID) Set @SGCount = (SELECT SGCount FROM dbo.Servers WHERE ServerID = @SID)
SET ROWCOUNT @top SELECT dbo.ExchangeMSData.*, dboExchangeMailStore.*, dbo.GetExchSGInfo(@SID,@SGCount) As ExchangeSG, dbo.Servers.* FROM dbo.Servers INNER JOIN ExchangeSG ON dbo.Servers.ServerID = ExchangeSG.ServerID INNER JOIN dbo.ExchangeMailStore ON ExchangeSG.StorageGroupID = dbo.ExchangeMailStore.StorageGroupID INNER JOIN dbo.ExchangeMSData ON dbo.ExchangeMailStore.MailstoreID = dbo.ExchangeMSData.MailstoreID WHERE (dbo.Servers.ServerName = @ServerName) ORDER BY dbo.ExchangeMSData.[TimeStamp] DESC, dbo.ExchangeSG.[TimeStamp] DESC
SET ROWCOUNT 0
And the Function: SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO
CREATE FUNCTION [dbo].[GetExchSGInfo] ( @SID INT, @SGCount INT ) RETURNS TABLE AS RETURN ( SELECT TOP (@SGCount) * FROM dbo.ExchangeSG WHERE ServerID = @SID ORDER BY [TimeStamp] )
I am a bit confused by the difference between a stored procedure and a table-valued function. Can somebody please either give me a simple explanation, or point me at something I can read.
I thought I had it worked out, and had coded some action queries as stored procedures, and I wrote a table-valued function that was effectively an encapsulated SELECT so that SELECT * FROM Spouse(@ID) worked fine. Then I wanted to use a function SpousePair, that was similar to Spouse, to power a Gridview. I discovered that I couldn't. It seems that a SQLDataSource requires either a SELECT statement or a stored procedure. So I wrote a stored procedure SpousePair(@ID1, @ID2).
I find that whereas I tested Spouse with SELECT * FROM SPOUSE(@ID) I tested SpousePair with EXEC SpousePair @ID1 @id2
Now I want to combine these: if I could I would write SELECT * FROM SPOUSE(@ID) WHERE SPOUSEID NOT IN (SELECT SPOUSEID FROM SpousePair(@ID1, @ID2))
However this is invalid because you can't put a stored procedure in a Select statement, and SELECT .... NOT IN (EXEC SpousePair @ID1 @ID2) is also invalid.
Is there any alternative to creating a table-valued function, SpousePairA, that is identical to SpousePair but coded as a function. I'm reluctant to do this because then I'll have two bits of quite complicated SQL logic to maintain.
fix the below SP which is having cursor output with an alternate logic which can make the SP work.(may be using temp table or any other option)
CREATE PROCEDURE Get_IDS /* * SSMA warning messages: * O2SS0356: Conversion from NUMBER datatype can cause data loss. */ @p_uid float(53), @return_value_argument varchar(8000) OUTPUT
I'm getting syntax errors that just aren't helping me at all, so I thought maybe what I'm trying to do can't be done. I'm creating a UDF with 4 parameters, and I want it to return a result set (i.e. a table). But I want a different result set depending upon the value of one of the parameters. This works totally fine as a SP, but I can't tell where to put the RETURN clause(s) on the UDF. I've got: CREATE FUNCTION myFunction(@param1,...,@param4) RETURNS TABLE AS BEGIN IF @param1='x' BEGIN RETURN(SELECT columns FROM TableX) END ELSE BEGIN RETURN(SELECT columns FROM TableY) END END I get an "Incorrect syntax near the keyword 'IF'" error. I also tried using just one RETURN() wrapped around the outside of the IF construction (right after the very first BEGIN and before the last END), but to no avail. I get no errors when I run this logic as an SP. Is this type of construct not allowed in a UDF? Is there an alternative? I can't just leave this as a proc because I'm going to have to call these results from several views. Help!
I'm probably just being thick but is there a way to use conditional statements within aggregate functions? I'm trying to do something along the lines of the following -
Code:
SELECT SUM( CASE WHEN Currency='GBP' THEN TotalAmountCharged ELSE TotalAmountCharged/1.45 ) as total from bookshop_orders where year(OrderDate) = 2004 and month(OrderDate) = 9
My overarching goal is to generate sets of random Symptom records for each Enrollee in a drug study, so that for each cycle (period of time), the code will insert a random number of random records for each enrollee.
I'm trying to return a number of random records from a table, but inside a table-valued function... (which could be my problem).
CREATE FUNCTION dbo.ufn_GetTopSymptoms ( @enrollID INT , @CTCVersion VARCHAR(20) , @NumRecords INT ) RETURNS TABLE
[Code] ....
But that ORDER BY NEWID() clause is illegal apparently, because here's the error it throws:
Msg 443, Level 16, State 1, Procedure ufn_GetTopSymptoms, Line 13 Invalid use of a side-effecting operator 'newid' within a function.
I was hoping I could return a set of enrollmentIDs and then use CROSS APPLY to generate a random set of records for each enrollmentID... is this not possible with APPLY? I was trying to avoid using a cursor...
The idea is basically to create all the Symptom records for all the patients in treatment cycle at once by using Enrollee OUTER APPLY dbo.ufn_GetTopSymtoms(dbo.Enrollment.EnrolleeID)
but that's clearly not working. Is there a way to do this without resorting to a cursor?
I have tables and a function as representated by the code below. The names for objects here are just for representation and not the actual names of objects. Table RDTEST may have one or multiple values for RD for each PID. So the function GIVERD will return one or multiple values of RD for each value of PID passed to it.
When I run the following query, I get the required result except the rows for CID 500 for which PID is NULL in table T1. I want the rows for CID 500 as well with PID values as NULL.
SELECT A.CID, A.ANI, A.PID, B.RD FROM T1 AS A CROSS APPLY GIVERD(A.PID) B
Hello, It is possible to write stored procedures which take table names as parameters; is it also possible to do this with table valued functions?
For example, a simple stored procedure is this:
CREATE PROCEDURE SelectTop(@tableName sysname) AS BEGIN
Execute('Select top 10 * from ' + @tableName + ';')
END
I want to be able to do the analogous thing with a table valued function (so that I can query the result set, without having to create a temp table). How should I do this (i.e., pass a tablename as an argument to a table valued function)?
