Citi Credit Card OFX Import Nightmare

[Brendan Coupe]

Running this command on the QFX (probably OFX) file zeros out the last
4 digits of the FITID value.

sed -ri 's/(\<FITID\>.*)[0-9][0-9][0-9][0-9]/\10000/' CitiVisa.qfx

When importing nothing matches so I deleted the old transactions in
KMM and hopefully this will prevent future duplication. Not sure if
duplicate FITID numbers will cause any other issues since all
transactions that happen on the same date now have the same FITID
number.

I have always had a similar problem with one of my investment banks
and now it's happening with the other one. Since they use direct
connect it's not as easy to fix. Shortening the download period to 3
days helps but it also means I can easily miss transactions.

Thomas, please remind my how to save the ofx file that is downloaded
with direct connect in KMM 5. I want to compare the files to see if
it's the same issue.

----
Brendan Coupe


----
Brendan Coupe


On Thu, Mar 28, 2019 at 11:57 AM Jeff Barlow <jeff at wb6csv.net> wrote:
>
> On 3/26/19 9:59 AM, Brendan Coupe wrote:
> > That would not surprise me since I switched to KMM 5 around the same time.
> >
> > A simple test is to download 2 oxf files from Citi. One with the current
> > billing cycle and one that has a shorter range with fewer transactions.
> > If you don't get duplicates then verify that the same transaction in
> > each has a different FITID number (diff -y). It might help to also see
> > what KMM is using as the bankid for the same transaction.
> >
> > Thanks for your data point. When is the last time you tried direct
> > connect? I'm hoping they fix this but haven't tried in a couple of months.
> >
> > ----
> > Brendan Coupe
>
> Okay... I tried direct connect... still broken.
>
> Also after tinkering a bit with the date range of the download I'm
> seeing the same sort of transaction matching issue that you are. It
> looks to me like Citi is generating the FITID numbers on the fly such
> that they are unique only within a given download. I wasn't seeing the
> issue at first only because I was repeatedly looking at the same
> (default) date range.
>
> --
> Later,
> Jeff