Automated reconciliation with GnuCash

[Gabriel M. Beddingfield]

Neil Williams wrote:

> No. Just click on the reconcile column in the register window and change

OK.  Thanks for the reply.

> GnuCash only supports QIF for importing new data currently. You'd need to

I've resigned myself to this... but I still wanted to check.

> If you import the data from QIF in the first place, what kind of check is
> it if you receconcile against the same QIF file??

I don't.  We enter transactions by hand from receipts.  So currently, we
don't really use QIF imports at all.  (Actually I do... but here's how:  My
wife enters her receipts in an Excel spreadsheet.  I export it to QIF and
import into gnucash.  This way she can enter data without having to use
linux.  She *really* likes Windows.)

> ?? That is a sweeping generalisation that is very hard to achieve in
> practice. What you consider easy to match does not translate well into
> code. You can't rely on the date of entry, you can't rely on matching the

I disagree.  In fact, GnuCash already does this when you import a QIF file. 
One of the steps in importing is to look for "potentially duplicate
transactions."  You affirm or deny whether it's actually a duplicate. 
That's all I want.

It's not hard to have a logic that:

(1) first compares amounts
(2) compares dates, say a radius of 7-14 days.
(3) compare descriptions for any matching strings
(4) ask the user for confirmation.

M$ Money already has this feature and it is the *ONLY* thing good about that
piece of crap software.  One of the down sides of Money's implementation is
that (IIRC) you don't have a chance to review its matches.

> You can import the QIF but you still need to reconcile the data from the
> import - it's an essential bit of maintenance and user involvement - you
> need to rely on this data, it's worth checking it's correct rather than
> trusting completely in some algorithm.

Nobody suggested trusting completely in an algorithm.  I just want a "sort"
algorithm... and then I go down the column and "approve" the matches and
figure out what to do when no match is found. I completely agree with you
about the importance of checking.  I just disagree about the importance of
spending an hour to do it.

Computers are good at sorting.
Humans are good at comparing abstract data.
Why not use both to do what they're good at?

-- 
                     G a b r i e l   M .   B e d d i n g f i e l d