adequacy of pyfproject to the processing of financial data for reporting purpose

[unccvince]

Hello,

I have successfully installed pyfproject and visited the tutorial for
first practice. Pyfproject is impressive at first look.

I'm not a developer, so my questions below may seem naive and the
description of my problem incomplete, sorry.

Below are two questions. I set the context first.

I have a project with 6-7 csv input tables that are joined. Some
tables are manually created by the user outside the ERP system. I
don't have a say on the way the data is exported from the ERP and
cannot ask for new queries to be written that would help me reduce the
number of CSV input tables.

1) The first processing must check the coherence of the data in its
entirety and output an error if it discovers an incoherence in the
dataset. The error returned qtops the processing and must be
comprehensible to the user so he may change/add/delete a value in his
table and reload the table

2) The second processing must check some functionnal rules and change
some data on the fly if necessary (i.e. some workhours amount was
affected to a wrong account and if some properties match a set of
rules, then the wrong account is replaced with the right account that
respects the business rule). This error correction must be reported in
a csv output report so that the comptrollers can ask their accountants
to do a better job of following the rules.

3) Then, the data is processed (grouped by sums on euros or
workhours).

4) Then the data is output in csv (approximately 20 reports).

5) The comptroller imports the output CSVs in spreadsheets and applies
the cosmetics.

question #1: with this short description, do you know whether
pyfproject can help me do this kind of stuff?

question#2: if so, does someone in France offer training services on
pyfproject?

Vince

[Florent Aide]

Vince,

That's exactly the kind of thing PyF does on a daily basis... and Yes
I could contact me & Jon in private email for training and/or services
in France.

I'll contact Jon on cell phone so he can answer you in this ML and you
get his email.

Best regards,
Florent Aide

[Jonathan Schemoul]

Hi,

I think your process is totally doable with PyF (even if some things
could be changed in your process to suit the dataflow system a bit):
- Do the checks while you process the data, and stop without outputing
csvs if there are errors
- The rest is ok with the dataflow way of working, but you may have to
do a little app for the "controller" (to load in spreadsheet and
things like this).

About training in France, I'll send you an email in private.

Regards,
Jonathan

On Wed, Feb 9, 2011 at 10:47 AM, unccvince <uncc...@gmail.com> wrote: