Hi Eran,
It is possible to run a mysql client, but it takes a bit of work. MySQL provide a pure python connector here http://dev.mysql.com/doc/refman/5.5/en/connector-python.html
It is licenced under GPL2 so you get the source. I got it to work by first installing into a regular Python 2.7 system on a linux box. Once installed and tested on the 'donor' find the mysql directory in site packages and copy the contents over to some other place for editing.
In the copy directory move everything into the toplevel directory, delete the empty init.py files and any .pyc files. Remove the sub directories. In your new mysql directory you should have 14 or so python files. You will need to change some filenames to something temporary to avoid over-writing.
To get it to work in pythonista start editing roughly as follows. Name the core init.py file to something like mysqldb (note: exclude .py otherwise pythonista won't import). Rename all the other files by just removing the .py extension. Rename the init.py file from the locales directory to locales.
Now for each file look for the 'mysql.connector...' import statements and remove the package path, just leave the core module which will have the same name as one of the above files. In the locales file edit line 48 (ish) which is an import statement and remove the path, leave the 'client_error' bit - this is the eng locale taken care of.
The next stage is to get the whole directory into pythonista which I ended up doing by importing the files in another app and re-creating them in pythonista using the clipboard - there are probably much more efficient ways to do this !
You should now have a mysql sub-directory in your main project directory. The last step in your main program is to import sys and insert the sub directory into sys.path - I just grab the last entry in path which is the project directory and do an os.path.join.
Longwinded, but it works for me. It does take a few seconds to load and connect on my ipad2, but not so long as to be tedious. I have been accessing various Mysql 5 databases for checking data and writing data - so far has worked exactly as a regular installation, and quite quick once loaded even for fairly large queries.
Hope this helps.