Is there a speed difference between python2.7 + mysqldb and python3.1 and myconnpy?

Hi Geert,

I am not sure where to post this, so I post it as a question, even
though it is more an answer ;)

We have a python3 based API with mysql backend using myconnpy...

And now we ported this to python2.7 (which was actually not that hard) and adapted our db-routines to mysqldb 1.2.3 (same functionality, same complexity of code, just mostly different syntax...).

Which enabled us to do a direct real life speed comparison :)

In a nutshell: python 3.1 with myconnpy yields the SAME speed than python 2.7 with mysqldb!

So if you can even optimize more we might end up getting even FASTER then mysqldb ;>

Here are the results:

Test: Select, modify, save one ~26k record 100 times.
Python 2.7 Server:
real 2m37.317s
user 0m0.450s
sys 0m0.120s

Python 3.1 Server:
real 2m33.375s
user 0m0.370s
sys 0m0.200s

Test: Select ~200 rows of ~200 bytes in length from one table, and
insert them into another table.
PYTHON 2.7 Server:
$ time python3.1 test4.py

real 0m4.354s
user 0m0.196s
sys 0m0.036s

$ time python3.1 test4.py

real 0m4.373s
user 0m0.196s
sys 0m0.048s

PYTHON 3.1 Server:
$ time python3.1 test4.py

real 0m4.400s
user 0m0.204s
sys 0m0.036s

$ time python3.1 test4.py

real 0m4.419s
user 0m0.200s
sys 0m0.040s

Test: Select ~1000 rows of ~25k bytes in length.
Python 2.7 Server:
$ time python3.1 test5.py
Got 1045 log-entries...

real 0m14.153s
user 0m7.512s
sys 0m0.416s

$ time python3.1 test5.py
Got 1045 log-entries...

real 0m13.987s
user 0m7.480s
sys 0m0.416s

Python 3.1 Server:
$ time python3.1 test5.py
Got 1045 log-entries...

real 0m14.116s
user 0m7.284s
sys 0m0.404s

$ time python3.1 test5.py
Got 1045 log-entries...

real 0m15.286s
user 0m7.832s
sys 0m0.452s

Cheers,

Knuth