database is a mix of myISAM and innoDB and collations

[ceberlin]

I find unusal that the PW created database turns out to be a mix of MyISAM and InnoDB tables,

Most collations are utf8_general_ci, but I fould also ascii_general_ci and _latin1_swedish_ci.

Maybe this is coming from modules...

I am not a database pro, I am designer, but to me this looks like potential trouble with searches, sortings and so on.

[ryan]

The only InnoDB table I'm aware of in ProcessWire comes from the DB session storage module, which is able to perform better with InnoDB than MyISAM. ProcessWire uses ascii_general_ci in some cases where the field can only ever support characters from the ASCII set, like "name" fields for instance. Everything else in ProcessWire uses utf8_general_ci. If you are seeing latin1_swedish_ci, chances are it came from an old version of ProcessWire or from a module that didn't specify collation. In some cases it matters, in others not. You may even want to change the collation in order to affect the way it sorts, for example. Ultimately I don't think you need to worry about it unless you have something that doesn't work properly or you want to change away from the default behavior of something. The defaults that we've chosen are what you want at least 99% of the time. 

[ceberlin]

Old modules from the past, I see.

We have some problem here with the find() and I wanted to be sure to rule out some possibilities.

[ryan]

What is the problem you are seeing with find() ?