sqlite3 assumed in python example script account_analysis.py ?

[Geert Janssens]

On Friday 28 January 2011, Derek Atkins wrote:
> Mark Jenkins <mark at parit.ca> writes:
> >> There is a comment in the script that says it only supports sqlite3
> >> files. The way the session is opened does indeed limit the choice of
> >> input files to sqlite3.
> >> 
> >> I'm just wondering, why was this done ?
> > 
> > Only to make invoking it on the command line more easy at the time I
> > wrote it.
> > 
> >> Simply changing the line that opens
> >> the session from
> >> 
> >>     gnucash_session = Session("sqlite3://%s" % gnucash_file,
> >>     is_new=False)
> >>  
> >>  to
> >>  
> >>     gnucash_session = Session(gnucash_file, is_new=False)
> >> 
> >> would allow the script to open any kind of database backend.
> > 
> > This would require the user to provide the filetype (xml:// or
> > sqlite3://) on the command line as well.
> > 
> > I don't have a problem with doing it that way either, it was just
> > laziness when I wrote it.
> 
> Um, "file://" should be understood, and "file://" ->
> "xml://","sqlite://" should be runtime-detected for opening existing
> files.  I.e. you should be able to run "Session("/path/to/file",
> is_new=False)" and it should auto-detect xml v. sqlite.
> 
> To put another way, there /should/ be no requirement that a type be
> applied.  The "file" type should be assumed, and the file loader should
> be able to detect xml v. sqlite.
> 
> If this is not the case then it is a bug (IMNSHO).
> 
> -derek
Indeed. And when I was working on that part of GnuCash I made sure it is 
indeed so.

Tested it yesterday with the python sample scripts by the way and can confirm 
that all works as intended.

Geert