Musings on ugly bank & credit card OFX data

My last vestige of a connection to Quicken is that I still have a Quicken
Credit card from Chase.  I think I used to be able to do direct downloads,
but that stopped a while back and I'm going to switch over to a credit card
from the Digital (Equipment Corp) Credit Union (that outlasted DEC).

Unfortunately, the DCU doesn't support direct downloads, so I'm
reduced to downloading OFX files.  Worse, the entries are like:

    <STMTTRN>
      <TRNTYPE>DEBIT
      <DTPOSTED>20120523000000[-5:EST]
      <TRNAMT>-40.40
      <FITID>2

      <NAME>PURCHASE  


      <MEMO>05-21-12 GULF OIL 91802719 CONCORD NH auth# 455599

    </STMTTRN>

<NAME> goes into the Description field, and <MEMO> goes into the memo
field.  Which is ugly.  And useless.  If <MEMO> went into the Description
field, it would still be ugly.  At least Chase offers up "<MEMO>GULF OIL
91802719", better but still shouting.

In the DCU's checking account, all the <NAME>s are "WITHDRAWL".  Which is
just as ugly.  And useless.  To its credit, MD manages to do a decent job
coming up with the right category for the transactions.  It could be worse!

I'm thinking of writing a Python hack to reformat these files and
have generic and specific matches on the MEMO field and put meaningful
stuff in the Description field before importing the file.  I also want
to see if I can use a command line invocation to trigger the import,
which would save that multi-step menu operation.

It will never be product quality code, but might be usable for other
folks.  Has anyone done something like this already?  Are there any
Python duffers who'd be interested in it?

  -Ric
  http://wermenh.com/
_______________________________________________
For Moneydance customer support, please visit http://help.infinitekind.com
[email protected]
http://moneydance.com/mailman/listinfo/moneydance-info