whatsnew 2.4 about itertools.group by:

[G?nter Jantzen]

In the documentation
http://www.python.org/dev/doc/devel/whatsnew/node7.html is written
about itertools.group by:

"""Like it SQL counterpart, groupby() is typically used with sorted
input."""

In SQL queries is the groupby clause not related to 'input order'.
This notion makes not much sense in SQL context.
SQL is based on relational Algebra. A SQL- table is based on an
unordered set of rows (implementation can be different, of course).

So the analogon of
----------------------

import itertools
L = [2,4,6, 7,8,9,11, 12, 14]
for key_val, it in itertools.group by(L, lambda x: x % 2): .... print key_val, list(it)
....
0 [2, 4, 6]
1 [7]
0 [8]
1 [9, 11]
0 [12, 14]

------------------------

Say you have a table 'example' with only one column 'i'

_______________ __________
select * from example;

I
----
2
14
6
7
8
9
11
12
4
_______________ ____________

the order of rows is not defined

Then you can group this table

_______________ _______________ ______________
select count(i), mod(i,2) from example group by mod(i,2)
COUNT(I) | MOD(I,2)
---------+---------
6 | 0
3 | 1
_______________ _______________ _____________

The result dos not depend on 'input order' or 'runs'