[GNC] Import Transaction Matcher issue in v4

Fross, Michael michael at fross.org
Mon Jul 27 21:56:49 EDT 2020


Hello David,

In looking at the QFX file, I only see one side of the transaction.  This
file has 3 distinct transactions (<STMTTRN> records) that are not related
and only one account ID is listed.

Michael

On Mon, Jul 27, 2020 at 5:56 PM David Carlson <david.carlson.417 at gmail.com>
wrote:

> There is another use case which GnuCash also needs to handle.  Some OFX
> files may contain both sides of an inter-account transfer between two
> accounts within the same bank because the OFX file can include multiple
> accounts.  In fact, I do that on a regular basis at one of my banks.  I do
> not test GnuCash's handling of these because I want to get the separate
> notes associated with each side and merge them together into the same
> transaction in my data file.  Thus I deliberately call both of them New
> rather than try to match them within the import process.  I have not
> checked to see if they have unique FITID's, but they would be under
> different ACCOUNTID's  and both can be accepted as New in release 2.6.19
> and prior, which is the latest version that I use.  I suppose ideally I
> would like gnuCash to detect this case and automatically apply the
> corresponding notes to both sides of the  transfer, so I don't have to do
> it manually.
>
> David Carlson
>
> On Mon, Jul 27, 2020 at 5:23 PM Fross, Michael <michael at fross.org> wrote:
>
>> Got it. Thanks.  Now I understand why it used to work.  And I certainly
>> wouldn't expect GnuCash to correct a bank's horrible behavior.  I'll keep
>> an eye on the FITID and see if that's what they are doing.
>>
>> Thanks everyone for your help.  Much appreciated.
>>
>> Michael
>>
>> On Mon, Jul 27, 2020 at 5:13 PM Jean Laroche <ripngo at gmail.com> wrote:
>>
>> > Ah, OK I get it now!
>> > Yes, you should check that from one day to the next, the FTID returned
>> > by citibank remains the same (for the same transaction). If not, then
>> > that's going to be a problem.
>> > What has changed recently with GC is that it not will not match a new
>> > imported transaction with one from the register that has an online id.
>> > Previous versions allowed that, so in your case, it's quite possible
>> > that from one day to the next, the same transaction would be given a
>> > different ID, but GC matched it to the same register transaction that
>> > was previously matched.
>> > So, on day 1, you'd match OFX1->Reg1, and on day two, you'd still match
>> > OFX2 -> reg1, even though reg1 had already been matched to OFX1 the
>> > previous day. So you'd see OFX2 appear as "match" and not "Add".
>> >
>> > The correct behavior is that OFX2 actually has the same ID as OFX1, and
>> > it's entirely skipped by GC. But it seems that your bank is messing
>> > their online ID, and this is what's causing the issue for you, along
>> > with the new GC behavior.
>> >
>> > I'm not sure how to fix your problem. It's clearly a bank issue but
>> > that's not terribly helpful as far as your problem goes. We should
>> > probably add an option to defeat the new GC import behavior,
>> > specifically for users like you whose bank send unreliable FTIDs/
>> >
>> > FYI, the new GC behavior was added because the old behavior was causing
>> > issues with people who had recurring daily transactions: imagine buying
>> > a cup of coffee every day: in that case GC will match all daily cups of
>> > coffee to the same register transaction (within a date range, of course)
>> > instead of adding the new one every day (which really is the correct way
>> > to do it).
>> >
>> >
>> >
>> > Jean
>> >
>> >
>> > On 7/27/20 2:45 PM, Fross, Michael wrote:
>> > > Hi Jean.  Let me try to better explain my issue.
>> > >
>> > > On most days, I download QFX files from my Checking and Credit Card
>> > > accounts and import them.  In this way I clear transactions and I can
>> > > see what's going on.  The next day I'll do so again, and the importer
>> > > skips those that have already been matched per your comment.s  I've
>> been
>> > > doing this for years and it's worked well.
>> > >
>> > > Since I updated to V4 the transactions that would have matched
>> > > previously now come up with "Match Missing!" in the importer.  I don't
>> > > want to add them as they would create duplicates.  Transactions that
>> are
>> > > not cleared seem to be able to be matched.  But those that are
>> > > previously cleared get the "Match Missing!" issue.
>> > >
>> > > Now, based on what you told me, the QFX file for a problem transaction
>> > > has a FITID that is one higher than the ONLINE_ID in the gnucash file
>> > > for the same transaction.  So the Match Missing items seems to be
>> > correct.
>> > >
>> > > If I duplicate the transaction, delete the original, it matches which
>> I
>> > > assume is because there is not an ONLINE_ID associated with the new
>> one.
>> > >
>> > > What I don't understand is if there is an issue with GNUCash, which is
>> > > why I asked if others are having the same issue, or if it is/was an
>> > > issue with CITIBANK QFX files.  If they changed the FITID for some
>> > > reason that would cause this.  Since no one else has said they are
>> > > having the issue, I'm assuming this is not a GNUCash issue.
>> > >
>> > > Does that better explain the issue?
>> > >
>> > > I really appreciate the time to respond and help me through this.
>> > >
>> > > Michael
>> > >
>> > >
>> > >
>> > >
>> > >
>> > >
>> > > On Mon, Jul 27, 2020 at 3:50 PM Jean Laroche <ripngo at gmail.com
>> > > <mailto:ripngo at gmail.com>> wrote:
>> > >
>> > >     There's a misunderstanding here.
>> > >     This is what GC does:
>> > >     - It looks at the OFX transaction, which always has an FITID.
>> > >     - IF GC sees this FITID in one of the register transactions, it
>> skips
>> > >     the OFX transaction, because it's already been imported (when a
>> > >     transaction gets imported from the ofx, the online_id is copied
>> into
>> > >     the
>> > >     register transaction, that's how GC remembers it's already
>> imported
>> > the
>> > >     transaction).
>> > >     - If the FITID is not found in the register, GC assumes that this
>> is
>> > a
>> > >     new transaction and attempts to match it with existing
>> transactions
>> > in
>> > >     your register. BUT:
>> > >          . It only looks at register transactions that have NO FITID
>> > (i.e.,
>> > >     register transactions that you entered manually, typically)
>> > >          . Among those, it looks for transactions that match pretty
>> > closely
>> > >     for the date and the amount.
>> > >          . It picks the one with the highest matching score, if that
>> > >     score is
>> > >     above a user-adjustable threshold.
>> > >
>> > >     So in the case you outline below, if the register transaction has
>> the
>> > >     same FITID as your ofx transaction, the ofx transaction will NOT
>> be
>> > >     imported (it will be just skipped).
>> > >
>> > >     I think I don't quite understand the problem you're having. Is it
>> > that
>> > >     you're importing OFX transaction and they're not matching register
>> > >     transactions that you entered manually?
>> > >     Or is it that they're being imported despite the fact they were
>> > >     imported
>> > >     previously?
>> > >
>> > >     J.
>> > >
>> > >     On 7/27/20 1:31 PM, Fross, Michael wrote:
>> > >      > Thanks Jean / John for your thoughts.  There is a register
>> entry
>> > >     that
>> > >      > matches, IMHO, very closely.   I increased the Match Display
>> > >     Threshold
>> > >      > from 1 to 3, and then to 6 (which appears to be the highest
>> value
>> > >      > allowed.)  Every transaction from the import says "Match
>> Missing."
>> > >      >
>> > >      > Digging around a bit, for the transaction in question, the QFX
>> > file
>> > >      > contains the FITID of 202007210003:
>> > >      >
>> > >      > <STMTTRN>
>> > >      > <TRNTYPE>CREDIT
>> > >      > <DTPOSTED>20200721120000
>> > >      > <TRNAMT>54.00
>> > >      > <FITID>202007210003
>> > >      > <NAME>ACH Electronic Credit
>> > >      > <MEMO>Expenses
>> > >      > </STMTTRN>
>> > >      >
>> > >      > My GNUCash file contains, for the same transaction has the
>> online
>> > id
>> > >      > being 202007210002
>> > >      >          <split:slots>
>> > >      >            <slot>
>> > >      >              <slot:key>online_id</slot:key>
>> > >      >              <slot:value
>> type="string">202007210002</slot:value>
>> > >      >            </slot>
>> > >      >
>> > >      > The online_ID is ...002 instead of ...003.  Changing the QFX
>> file
>> > to
>> > >      > match the online_id value seemed to work.  Now my question is
>> why
>> > >     would
>> > >      > this be different for *lots* of transactions.  Everything
>> worked
>> > >      > normally in v3, but this would not have changed as part of the
>> > >     release.
>> > >      > I'll check a few more problem transactions and see if I can
>> > detect a
>> > >      > pattern.  Perhaps Citibank is paying games....
>> > >      >
>> > >      > Michael
>> > >      >
>> > >      > On Sun, Jul 26, 2020 at 4:57 PM jean laroche <ripngo at gmail.com
>> > >     <mailto:ripngo at gmail.com>
>> > >      > <mailto:ripngo at gmail.com <mailto:ripngo at gmail.com>>> wrote:
>> > >      >
>> > >      >     To get a match you have to have a transaction in the
>> register
>> > >     that's
>> > >      >     sufficiently similar to the one you're importing, and that
>> > >     has not been
>> > >      >     imported/matched before.
>> > >      >     In your case, it could be one of these reasons (I can't see
>> > >     the image):
>> > >      >     - There's no matching transaction in your register (no
>> > existing
>> > >      >     transaction has amount close, and a date close to the
>> > >     imported one)
>> > >      >     - There's a matching transaction but it's already been
>> > >     matched to an
>> > >      >     imported transaction at some point so it's not available to
>> > >     be matched
>> > >      >     to the new imported one.
>> > >      >     - There's a matching transaction that's available, but the
>> > >     match score
>> > >      >     is below the threshold that allows the transaction to be
>> > >     shown as a
>> > >      >     potential match. Too large a date mismatch can cause that.
>> > >      >
>> > >      >     Can you check whether you're in one of these 3 cases? If
>> > >     you're in case
>> > >      >     3, you can lower the minimum matching threshold in the
>> > >     preferences and
>> > >      >     see if that helps.
>> > >      >     J.
>> > >      >
>> > >      >
>> > >      >     On 7/26/2020 2:44 PM, John Ralls wrote:
>> > >      >      > If there's no matching transaction already in the
>> account
>> > then
>> > >      >     there's nothing to clear. In that case only adding or not
>> > >     makes sense.
>> > >      >      >
>> > >      >      > Regards,
>> > >      >      > John Ralls
>> > >      >      >
>> > >      >      >
>> > >      >      >> On Jul 26, 2020, at 1:56 PM, Fross, Michael
>> > >     <michael at fross.org <mailto:michael at fross.org>
>> > >      >     <mailto:michael at fross.org <mailto:michael at fross.org>>>
>> wrote:
>> > >      >      >>
>> > >      >      >> Hello all,
>> > >      >      >>
>> > >      >      >> I sent this earlier this month and didn't see any reply
>> > so I
>> > >      >     thought I
>> > >      >      >> would try again.   Has anyone else seen these issues?
>> I
>> > use
>> > >      >     Citibank and
>> > >      >      >> perhaps it's a Citibank issue, but I did not have this
>> > >     problem
>> > >      >     on v2 or v3.
>> > >      >      >>
>> > >      >      >> Thanks all.  I appreciate the help.
>> > >      >      >>
>> > >      >      >> Michael
>> > >      >      >>
>> > >      >      >>
>> > >      >      >> On Sat, Jul 4, 2020 at 9:48 AM Fross, Michael
>> > >     <michael at fross.org <mailto:michael at fross.org>
>> > >      >     <mailto:michael at fross.org <mailto:michael at fross.org>>>
>> wrote:
>> > >      >      >>
>> > >      >      >>> Hello all,
>> > >      >      >>>
>> > >      >      >>> I typically download QFX files from my banks every day
>> > >     or two,
>> > >      >     import them
>> > >      >      >>> to clear them in Gnucash.  Worked great.  However,
>> ever
>> > >     since
>> > >      >     upgrading to
>> > >      >      >>> v4, the importer seems to have trouble matching.  Most
>> > >     of the
>> > >      >     imported
>> > >      >      >>> transactions are listed in the importer as (A)dd, but
>> > >     when I select
>> > >      >      >>> (C)lear for them it says match missing.
>> > >      >      >>>
>> > >      >      >>> This has occurred for several accounts.  Here is a
>> simple
>> > >      >     credit card
>> > >      >      >>> example, although for my checking account, there are
>> > dozens
>> > >      >     like this.  The
>> > >      >      >>> top portion shows the register with the Sprint bill
>> > >     cleared.
>> > >      >     The date,
>> > >      >      >>> amount, and name (mostly) match.
>> > >      >      >>>
>> > >      >      >>> [image: image.png]
>> > >      >      >>>
>> > >      >      >>> Not sure if there is just something wrong with my
>> setup
>> > or
>> > >      >     not.  Perhaps a
>> > >      >      >>> bug?  Are others experiencing this?  Any ideas to get
>> the
>> > >      >     matcher matching
>> > >      >      >>> again?  Something need to get cleared out?
>> > >      >      >>>
>> > >      >      >>> For those of us in the US, happy Independence Day.
>> > >     Thank you
>> > >      >     all for your
>> > >      >      >>> assistance.
>> > >      >      >>>
>> > >      >      >>> Michael
>> > >      >      >>>
>> > >      >      >>
>> <image.png>_______________________________________________
>> > >      >      >> gnucash-user mailing list
>> > >      >      >> gnucash-user at gnucash.org
>> > >     <mailto:gnucash-user at gnucash.org> <mailto:
>> gnucash-user at gnucash.org
>> > >     <mailto:gnucash-user at gnucash.org>>
>> > >      >      >> To update your subscription preferences or to
>> unsubscribe:
>> > >      >      >>
>> https://lists.gnucash.org/mailman/listinfo/gnucash-user
>> > >      >      >> If you are using Nabble or Gmane, please see
>> > >      > https://wiki.gnucash.org/wiki/Mailing_Lists for more
>> information.
>> > >      >      >> -----
>> > >      >      >> Please remember to CC this list on all your replies.
>> > >      >      >> You can do this by using Reply-To-List or Reply-All.
>> > >      >      > _______________________________________________
>> > >      >      > gnucash-user mailing list
>> > >      >      > gnucash-user at gnucash.org <mailto:
>> gnucash-user at gnucash.org>
>> > >     <mailto:gnucash-user at gnucash.org <mailto:gnucash-user at gnucash.org
>> >>
>> > >      >      > To update your subscription preferences or to
>> unsubscribe:
>> > >      >      > https://lists.gnucash.org/mailman/listinfo/gnucash-user
>> > >      >      > If you are using Nabble or Gmane, please see
>> > >      > https://wiki.gnucash.org/wiki/Mailing_Lists for more
>> information.
>> > >      >      > -----
>> > >      >      > Please remember to CC this list on all your replies.
>> > >      >      > You can do this by using Reply-To-List or Reply-All.
>> > >      >
>> > >      >     _______________________________________________
>> > >      >     gnucash-user mailing list
>> > >      > gnucash-user at gnucash.org <mailto:gnucash-user at gnucash.org>
>> > >     <mailto:gnucash-user at gnucash.org <mailto:gnucash-user at gnucash.org
>> >>
>> > >      >     To update your subscription preferences or to unsubscribe:
>> > >      > https://lists.gnucash.org/mailman/listinfo/gnucash-user
>> > >      >     If you are using Nabble or Gmane, please see
>> > >      > https://wiki.gnucash.org/wiki/Mailing_Lists for more
>> information.
>> > >      >     -----
>> > >      >     Please remember to CC this list on all your replies.
>> > >      >     You can do this by using Reply-To-List or Reply-All.
>> > >      >
>> > >
>> >
>> _______________________________________________
>> gnucash-user mailing list
>> gnucash-user at gnucash.org
>> To update your subscription preferences or to unsubscribe:
>> https://lists.gnucash.org/mailman/listinfo/gnucash-user
>> If you are using Nabble or Gmane, please see
>> https://wiki.gnucash.org/wiki/Mailing_Lists for more information.
>> -----
>> Please remember to CC this list on all your replies.
>> You can do this by using Reply-To-List or Reply-All.
>>
>
>
> --
> David Carlson
>


More information about the gnucash-user mailing list