Installation guide

5. mxODBC Overview

This generates a result set with column COLUMN_TYPE which has the needed

information.

Please see the documentation in section 7.6.1 Catalog Methods for

cursor.procedurecolumns() for details.

5.6.4 Special constraints of some ODBC drivers

Mixing output parameters and output result sets

Some ODBC drivers, most notably, the MS SQL Server Native Client, send the

output parameter data only after any output result sets which the stored

procedure may have created.

Since the mxODBC cursor methods return the output parameter right after

executing the stored procedure and before fetching any result sets, the output

parameters are not yet updated. Furthermore, there is no method of accessing the

output parameters after those methods have returned.

As a result, you won't see output parameter updates in your application, if you

are using both output parameters and output result sets in your stored

procedures.

The work-around for this is to pass back the output parameter data using an

additional result set and then fetching this using the

cursor.nextset() method

after the other result sets. Also see section 5.6.5 Using Result Sets for passing

back Output Data on this topic.

Example:

cursor.callproc('procname', parameters)

# Get the original output result set

output_result_set = cursor.fetchall()

# Switch to the new output parameter result set

cursor.nextset()

output_parameter_result_set = cursor.fetchall()

Using None as value for output parameters

mxODBC uses the value passed to the cursor execution method as hint for the

output type of the output parameter, in case the ODBC driver does not provide

this information.

If you pass in

None as placeholder for the output parameter value, mxODBC is, in

some cases, unable to determine the correct output type. It then defaults to using

VARCHAR(4000) output parameter value and sends back the data as character

string.