Re: prob in sp - Database Discussion Boards

CREATE TABLE [Table2] (
[IDNO] [int] NULL ,
[RefID] [int] NULL ,
[Code] [varchar] (2) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
[dt] [smalldatetime] NULL
) ON [PRIMARY]
GO

CREATE TABLE [Table1] (
[IDNO] [int] NULL ,
[RefID] [int] NULL
) ON [PRIMARY]
GO

INSERT INTO Table1 (IDNO, RefID) VALUES (1, 100)
INSERT INTO Table1 (IDNO, RefID) VALUES (2, 200)
INSERT INTO Table1 (IDNO, RefID) VALUES (3, 300)
INSERT INTO Table1 (IDNO, RefID) VALUES (4, 400)
INSERT INTO Table1 (IDNO, RefID) VALUES (5, 500)

INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (1, 100,'a', '1/1/2005 1:01:00 AM')
INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (2, 100,'b', '1/1/2005 1:02:00 AM')
INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (3, 100,'c', '1/2/2005 2:03:00 AM')
INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (4, 100,'d', '1/2/2005 3:00:00 AM')
INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (5, 100,'f', '1/3/2005 1:00:00 AM')
INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (6, 200,'a', '1/1/2005 4:55:00 AM')
INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (7, 200,'c', '1/1/2005 4:59:00 AM')
INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (8, 200,'f', '1/1/2005 5:02:00 AM')
INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (9, 300,'a', '1/1/2005 1:00:00 AM')
INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (10,300,'c', '1/1/2005 2:00:00 AM')
INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (11,300,'d', '1/1/2005 3:00:00 AM')
INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (12,300,'f', '1/1/2005 4:00:00 AM')
INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (13,400,'a', '1/1/2005 1:00:00 AM')
INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (14,400,'c', '1/1/2005 2:00:00 AM')
INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (15,400,'e', '1/1/2005 3:00:00 AM')
INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (16,400,'ff', '1/1/2005 4:00:00 AM')
INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (17,500,'c', '1/1/2005 1:00:00 AM')
INSERT INTO Table2 (IDNO, RefID, Code, dt) VALUES (18,500,'f', '1/1/2005 2:00:00 AM')

The task changed. I don't care if the last record is 'f' anymore. I now only care if the second to last record is 'c'.

Why the change? Because as I started working on this problem I came to the conclusion that all the items in table1 will always have a last record of 'f'.

--Here is what I have so far.
--If you test against the test data it almost works.
--My only issue now is that I'm seeing 500 which is wrong
--because there are only 2 records not 3.
--See I'm trying Smile | :)

select * from table1
where refid in
(
select top 1 refid from
( select top 3 * from table2 where refid = table1.refid order by dt desc
) as tt where tt.code = 'c' order by dt
)

Thanks go out to Michael Potter who reminded me about correlated subqueries.
-Richard

I think you are mis-stating your criteria. 'c' is the 2nd to a last record using RefId 500. It is not the 2nd to last in RefId 100, 300, 400. Your code dumps all RefIds. Here is some code that reports if 'c' is the 2nd to last record (by the DateTime order that you didn't specify in the first message).

SELECT *
FROM Table1 
WHERE  
    (SELECT COUNT(*)
     FROM Table2
     WHERE RefId = Table1.refId AND
           dt > (SELECT dt
                 FROM Table2
                 WHERE RefId = Table1.refId AND
                       Code = 'c')) = 1

If you are truely looking for the 3rd to last then just change the '1' to '2'. If you want to make sure that 'c' can not be the first than add that criteria also.

Absolutely brilliant!!! I knew there had to be a better way to do it. Thank you very much for your assistance. I’ll attempt to implement this solution tomorrow. Hopefully my page will load much quicker now.

Again thank you. And I’m sorry if my initial remarks where not exactly what I was attempting to get. D'Oh! | :doh:

It’s almost like I became a manager or something for a minute... ahhhh... I’m better now Smile | :)

-Richard

i'm not an SQL guru but, you can create a View with this two tables and join the RefID of Table1 with the RefID of table2 this will give you (with the correct WHERE clause) what you want
hope this help you.

GanDad

i have created sp for insert and edit data in the table called
EPG_tblResourceMaster.

when i run my sp....
exec dbo.USP_AddEditResourceMaster '001','r002','veh','m001','s002','1000.0' it execute sucess fully.

but when i run the select * from dbo.EPG_tblResourceMaster before run the sp...and after again try to run my sp it gives error like....

Server: Msg 50000, Level 16, State 1, Procedure USP_AddEditResourceMaster, Line 76
Rowcount 988982068 inserting into Region

how can i solve this prob......

Thanks

Amal

Without seeing the code for the sp, I wouldn't have a clue.

Anger is the most impotent of passions. It effects nothing it goes about, and hurts the one who is possessed by it more than the one against whom it is directed.
Carl Sandburg

Actually nice though..... Big Grin | :-D

any way if u need my sp code here it is.but i guess thr is no prob in sp.any way it's upto u chk it and give me solution. D'Oh! | :doh:

SET QUOTED_IDENTIFIER ON
GO
SET ANSI_NULLS ON
GO

/*AMAL FERNANDO ON 7 FEB 2005 */

ALTER PROC [USP_AEDMasterCode]
(@szCompCode VARCHAR(5),
@szMasterCode VARCHAR(5),
@szDesc VARCHAR(20),
@szCategory VARCHAR(10),
@szCategoryID VARCHAR(2),
@opType VARCHAR(6),
@type VARCHAR(6) OUTPUT,
@errDesc VARCHAR(20) OUTPUT)
AS

DECLARE @@dtLastLogin DATETIME
DECLARE @error INT
DECLARE @rowcount INT
DECLARE @errorMsg NVARCHAR(512)
DECLARE @spName SYSNAME

SELECT @error = @@error,
@rowcount = @@rowcount,
@spName = Object_Name(@@ProcID)

SET NOCOUNT ON
BEGIN TRAN

IF @opType = 'DELALL'
BEGIN
DELETE FROM [DBEPGlobal].[dbo].[EPG_tblMasterCodes]
SET @type = 'DELALL'
END
ELSE IF @opType = 'DEL'
BEGIN
DELETE FROM [DBEPGlobal].[dbo].[EPG_tblMasterCodes]
WHERE MC_szCompCode = @szCompCode
AND MC_szMasterCode = @szMasterCode
AND MC_szCategoryID = @szCategoryID
SET @type = 'DEL'
END
ELSE
IF EXISTS(SELECT * FROM EPG_tblMasterCodes
WHERE MC_szCompCode = @szCompCode
AND MC_szMasterCode = @szMasterCode
AND MC_szCategoryID = @szCategoryID)
BEGIN
UPDATE [DBEPGlobal].[dbo].[EPG_tblMasterCodes]
SET [MC_szCompCode] = @szCompCode,
[MC_szMasterCode] = @szMasterCode,
[MC_szDesc] = @szDesc,
[MC_szCategory] = @szCategory,
[MC_szCategoryID] = @szCategoryID
WHERE MC_szCompCode = @szCompCode
AND MC_szMasterCode = @szMasterCode
AND MC_szCategoryID = @szCategoryID
SET @type = 'UPDATE'
END
ELSE
BEGIN
INSERT INTO EPG_tblMasterCodes
(MC_szCompCode,
MC_szMasterCode,
MC_szDesc,
MC_szCategory,
MC_szCategoryID)
VALUES(@szCompCode,
@szMasterCode,
@szDesc,
@szCategory,
@szCategoryID)
SET @type = 'SAVE'
END

IF @@ERROR <> 0
BEGIN
ROLLBACK TRAN
RAISERROR('Error %d inserting into Region', 10, 1,@spName, @error,@errorMsg)
SET @errDesc = @error
RETURN(@error)
END
ELSE IF @rowcount <> 1
BEGIN
ROLLBACK TRAN
RAISERROR('Rowcount %d inserting into Region', 10, 1,@spName, @rowcount,@errorMsg)
WITH SETERROR
SET @errDesc = @@error
RETURN(@@error)
END
ELSE
BEGIN
COMMIT TRAN
RETURN(0)
END

GO
SET QUOTED_IDENTIFIER OFF
GO
SET ANSI_NULLS ON
GO

Thanks Confused | :confused:

Your code sets a local variable @rowcount - different from @@ROWCOUNT - to whatever the last row count was when the procedure was called - i.e. how many rows were affected/returned by the previous statement executed on that connection.

If you need to check the row count or error code, you should use a SET statement just before you need to check it.

Stability. What an interesting concept. -- Chris Maunder

Dear mike thanks for your reply.any way i 'm new to sql server 2000 and i want follow best practises.so can u give me guide sp sample which handle all the eroors and return values.

Thanks

The scenario is that i have a webservice returning a strongly typed dataset to a client app with ostensibly the same typed dataset. Unfortunately they are not recognised as being of the same type and get the error message "cannot convert type 'CAClient.WSCA.AdminStaff to CAClient.AdminStaff'. CAClient is my client. WSCA is the reference to my webservice.

Calling Data from Webservice on The client:
currentUserInformation.myAdminStaff=(AdminStaff) m_WsData.GetAdminStaff(myTicket,currentUserInformation.userGroupId);
webservice:
[WebMethod]
AdminStaff typedDataSet = new AdminStaff();
Param = new SqlParameter("@id",Id);
string [] tableList= {typedDataSet.Tables[0].TableName};

SqlHelper.FillDataset(dbConn, CommandType.StoredProcedure," Confused | :confused:

GetStaff",typedDataSet,tableList, Param );
return typedDataSet;

dazzler

Im replying to my own cry for help as i have managed to resolve it. I think this is a problem that only newbies like myself are likely to encounter. It stems entirely from a lack of understanding. What i had done was to create a Typed Dataset in a Webservice and then duplicate it entirely on my client. I then stupidly assumed that i could assign the typed dataset returned by a call to the webservice to the one created in my client. These are obviously recognised as two distinct types even though structurally they are the same.

The way to resolve this issue is to create a reference to the webservices typed dataset in the client and not to duplicate it in the client. (sensible really)

cheers

dazzler

dazzler

Dear guy's i'm using Enterprise library 2005 for my asp.net project.in there i want read return value of my Stored procedure.my how can i edit my code..

Database db=DatabaseFactory.CreateDatabase();
string strsqlcmd="USP_AddEditResourceMaster";
DBCommandWrapper dbcmdwr=db.GetStoredProcCommandWrapper(strsqlcmd);
dbcmdwr.AddInParameter("@szCompCode",DbType.String, Resource.EPGRScompcode);
dbcmdwr.AddInParameter("@szResourceID",DbType.String,
Resource.EPGRSresID);
dbcmdwr.AddInParameter("@szResName",DbType.String, Resource.EPGRSresName);
dbcmdwr.AddInParameter("@szResType",DbType.String,
Resource.EPGRSresType);
dbcmdwr.AddInParameter("@szResMesureType",DbType.String, Resource.EPGRSresMType);
dbcmdwr.AddInParameter("@dResMinCost",DbType.Double,
Resource.EPGRSMincost);
dbcmdwr.AddInParameter("@dResMaxCost",DbType.Double,Resource.EPGRSMaxcost)db.ExecuteNonQuery(dbcmdwr);

Thanks

Use dbcmdwr.AddParameter.
The 4th parameter is 'direction', and you use ParameterDirection.ReturnValue for the value. After the stored procedure is run, just get the value like you would any other output parameter:
dbcmdwr.GetParameterValue("@MyRetValParameter")

I don't use return values in my stored procedures because not all db support return values, so I code my stored procedures to use output parameters for all values returned. I think this is mentioned in the documentation for the enterise library.

Thanks Scott....now i knoe how to read return values.u mentioned about the out parameter and i don't know the advantages use outpara than para.any way now i'm use pot para's to read return values from sp.

Finally i have another prob with exception handling in Enterprise library.cas
in my application every time it gives common error called "access denied to registry and all that".can u give me a sample application how to handle exception in better way.

Thank you

Amal

I have not used the exception handling enterprise code yet, so I do not know about the error. I just started using the data enterprise application block about 2 weeks ago.

I get errors like these in the application log whenever I run code with the data application block:
"Failed to create instances of performance counter 'Total Command Executions' - The requested Performance Counter is not a custom counter, it has to be initialized as ReadOnly.."

I will eventually go through the code and find where this is happening in the enterprise application block code, but for now I just ignore it.

OK man,thanks for the reply.any way if get a solution pls let me know.

What happen or how can I prevent some error if i do an update to a table and there is nothing in the table to begin with?

In my code:

<br />
	try{<br />
		CString mySelect;<br />
		mySelect.Format("UPDATE TSK2 SET TSK2_DATE_TIME = CURRENT_TIMESTAMP");<br />
		m_pConnection->Execute(_bstr_t(mySelect),NULL, NULL);<br />
	}<br />
	catch (_com_error &e)<br />
	{<br />
		_bstr_t bstrError(e.ErrorMessage());<br />
		CString strError = (char*)bstrError;<br />
		AfxMessageBox(strError);<br />
	}<br />
	catch (...)<br />
	{<br />
		AfxMessageBox("Unknown Error!");<br />

Where m_pConnection is a pointer to my _ConnectionPtr. Will I get some thing back from "Execute" that will tell me, that I have to do an INSERT instead of an UPDATE if there is nothing in the table to begin with?

Also I got this section of code from the MSDN site:

<br />
     // Define temporary strings for output conversions.<br />
        // Initialize to first record's values.<br />
        _bstr_t bstrTitle;<br />
        _bstr_t bstrType;<br />
<br />
        // Enumerate Recordset and print from each.<br />
        while(!(pRstTemp->EndOfFile))<br />
            {<br />
            // Convert variant string to convertable string type.<br />
            bstrTitle = pRstTemp->Fields->GetItem("Title")->Value;<br />
            bstrType  = pRstTemp->Fields->GetItem("Type")->Value;<br />
            printf("\t%s, %s \n", <br />
                (LPCSTR) bstrTitle,<br />
                (LPCSTR) bstrType);<br />
    <br />
            pRstTemp->MoveNext();<br />

If I do not know what type of data that I'm receving from a table can I do what is being done in this code? Or will this only work for string values?

Here is the link to the MSDN website where I found this:
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/ado270/htm/mdmthexecutexvc.asp

Thanks

Tom Wright
tawright915@yahoo.com

The answer to your first question is no, it will gladly do nothing gracefully, and if there are any records in the table, it will try to update the TSK2_DATE_TIME field in all of them (failing if they have a unique index or constraint or ar key fields) since there is no "where" clause to select a unique record. I seem to recall that the Execute method does return a long containing the number of records affected, you could check this for 0...

The second should work for most data types, but might fail if the fields are null. There is a nice wrapper class for ADO Here[^] that takes some of the pain away...

Anger is the most impotent of passions. It effects nothing it goes about, and hurts the one who is possessed by it more than the one against whom it is directed.
Carl Sandburg

So basically there is no way of knowing whether to do an update (thinking that the table has something all ready in it) or to do an insert (thinking that the table is empty).

Here is what I'm thinking. When a person first creates a table, there is nothing in it. So after the table is created and a programmer goes and write a program that will update a record in the table, their program will error out since there is no record to update. So how would the programmer know whether to do an insert or not?

Thanks

Tom Wright
tawright915@yahoo.com

"Select Count(*) as Records from MYtable"

if Records == 0, then insert, else update.

Actually, I am baffeled that yo want to update records without knowing which record or how many records.

Since Update statements are intended for changing the values of some fields in some specific existing records, it doesn't usually make much sense to blindly update everything (all records), which is what your query does since it has no 'where' clause to selecect specific records to update. Your query won't generate an error if there are no records, it will just gracefully return a rows effected variable containing 0.
If there are 1 million rows, it will change them all if it can, again without returning an error.

If there is a question as toi whether or not a specific record exists, one would normalkly do something like:
strQuery = "If NOT Exists(select x from MyTable where x = 23) Inset into MyTable (x,y) values(23,'this record was missing') ELSE update Mytable set y ='the record was there' where x = 23"
Myconnection.Execute(strQuery....

An even better practice is to make queries like this into stored procedures with appropriate parameters, and execute the sp, passing ony the parameters needed (23 in this example)

Anger is the most impotent of passions. It effects nothing it goes about, and hurts the one who is possessed by it more than the one against whom it is directed.
Carl Sandburg

Well Rob. After looking at my post, I realized that and UPDATE is not what I wnated to do. This was before I read your last post. I realized that I wanted to do an insert.

I guess my question pertained to a thought I had. Here it is, when I create a new table, all fields are blank. Let say that there should be records in this table but they are not in there yet. (This is based on the fact that we do a lot of conversion from C-ISAM databases into SQL database) Should I always be checking to see if there are records in the table before I do an update? That was the question. Because if there where no records to update then it would error out.

I guess in my try, catch I could do a try update and if error catch exception and do insert. But that is assuming that it errored based on the idea that there where no records to update. Bad form. So hence my post, is there a way to determine when to do an update or an insert.

In my case I found out that i should always be doing an insert.

Thanks for the help

Tom Wright
tawright915@yahoo.com

Is there a way to capture the sql server messages? I have some reporting software that only works if rows are returned and I have some Sp that I want to run from this software that return no rows just server messages. I can turn off the check for rows in the software (just an if statement) but can’t figure out how to capture server messages.
Thanks for any help

Ronald Hahn, CNT - Computer Engineering Technologist

New Technologies Analyst

HahnTech Affiliated With Code Constructors
Edmonton, Alberta, Canada
Email: rhahn82@telus.net

http://www.588188.com/netbook/sqlserver2000/adosql/adoprg03_59pr.htm

I have a VB.net standalone application. I want a method of merging a table from the database application with a master database. Replication is not an option.

I have tried dumping the table from db1 into a dataset (dataset1), and then the same table from db2 into a dataset (dataset2). Will using the dataset merge method combine the data to enable me to update the master.

IE.
dataset1.Merge(dataset2)
dataset1.acceptchanges

dataAdapter1.Update(dataset1, "Table1")

It seems that dataset1 gets dataset2's data appended, but the datasource does not update.

Please advise a method to do this.