如何使用pyodbc获取SQL Server存储过程返回的行集？

Question

My question is in close connection to stackoverflow/9915788 .

I am using pyodbc to upload data to an external SQL database and for this I use a stored procedure of that database. This procedure returns two things:

• a rowset (a single row which contains the text "Success" if the procedure worked fine or the text "Fail" followed by any available reason if it failed)
• an integer value (0 for success, -1 for failure)

I managed to read out the integer value using the approach described in the link above, but I would like to read out the rowset as well because I would like to get further information in case of any errors.

Any ideas how to do this? I'm not a SQL expert, but in my understanding a rowset should be a "cursor like" object since it should be possible to loop that row; but how do I get this in a select statement?

This is how I get the integer value:

def CallStoredProc(conn, procName, *args):
    sql = """SET NOCOUNT ON;
            DECLARE @ret int
            EXEC @ret = %s %s
            SELECT @ret""" % (procName, ','.join(['?'] * len(args)))
    return conn.execute(sql, args).fetchall()

But I have no idea how to get the row, which should also be available somewhere..

Answer 1

For the following stored procedure in SQL Server

CREATE PROCEDURE [dbo].[IsItGord] 
    @p1 varchar(50) = 'everybody'
AS
BEGIN
    DECLARE @retval int;
    SET NOCOUNT ON;
    IF @p1 = 'Gord'
    BEGIN
        SET @retval = 0;
        SELECT 'Success' AS response;
    END
    ELSE
    BEGIN
        SET @retval = -1;
        SELECT 'Fail - ' + @p1 + ' is not Gord.' AS response;
    END
    RETURN @retval;
END

the following pyodbc code

cnxn = pyodbc.connect(connStr)
sql = """\
DECLARE @return_value int;
EXEC    @return_value = [dbo].[IsItGord] ?;
SELECT  'Return Value' = @return_value;
"""
crsr = cnxn.execute(sql, ("Gord"))
data = crsr.fetchall()
print(data)
crsr.nextset()
data = crsr.fetchall()
print(data)

prints

[('Success', )]
[(0, )]

and if I change the .execute() to

crsr = cnxn.execute(sql, ("Fred"))

it prints

[('Fail - Fred is not Gord.', )]
[(-1, )]

Answer 2

在SQL Server中执行存储过程之前，始终使用：set nocount ON。