Installation guide

5. mxODBC Overview

PostgreSQL

mxODBC defaults to forward only cursors, since the driver becomes unusable

with other settings.

IBM DB2

Just like MS SQL Server, IBM DB2 supports static cursors as well, so you can

enable these, if you need scrolling support or

.rowcount information.

There is a performance penalty of about 2x in using static cursors.

connection = mx.ODBC.Windows.DriverConnect(…)

connection.cursortype = mx.ODBC.Windows.SQL.CURSOR_STATIC

# All cursors created on connection will then default to static

# cursor type.

Note that with forward cursors, the .rowcount attribute does not always give

correct results with DB2.

5.9 Custom Cursor Row Objects and Row

Factory Functions

In some situation you may want to have mxODBC return a different Python object

type or class when fetching rows from the database, e.g. ones which offer not

only indexes based access to the row fields, but also named attribute or index

access.

mxODBC 3.3 introduced two new cursor attributes

cursor.row and

cursor.rowfactory which can be used to customize the objects returned by

mxODBC in result sets when using the

cursor.fetch*() methods.

Default is to return standard Python tuples for rows in the result, since this is the

most performant way of providing access to the data.

With the new attributes, it is possible to have mxODBC automatically return other

sequence types (e.g. lists), hybrids providing sequence/mapping/attribute access

to the columns or higher-level abstraction layer objects.

5.9.1 Cursor Row Constructor: cursor.row

In order to have mxODBC use a different row constructor than the tuple

constructor, set the

cursor.row attribute to a function or class which takes a

single parameter, the row tuple, and returns an object for the row.