Cliff, did you not see my reply or the one from Liz? Making the corrections *after* the import is never going to teach GnuCash how to match those transactions correctly. The corrections are to be done as part of the import matching process.
The documentation covers this. See the Help Guide "§6.15.6. Generic Transaction Import Matcher", and the Tutorial & Concepts Guide "§3.1. Importing Transactions from Files”. Liz asked for additional info on your system and version so more specific advice can be offered. Regards, Adrien p.s.—I’m copying you this time in case you aren’t subscribed to the list, but I’d advise it if you are not. If you send a reply, be sure to send it to the list. (or use reply-all) > On Jun 15, 2020 w25d167, at 2:30 PM, Cliff Kushler <ckush...@gmail.com> wrote: > > Hi Gio, > > Thanks for your quick response! > > Yes, I have been importing transactions for many years. However, it seems > that Gnucash is not getting any better - I have to correct transactions from > the same source over and over again, month after month. There is a window > that appears during the import process (before the transactions are posted in > the actual account registers) that shows all of the transactions to be > imported, with check boxes labelled “A” (accept? not sure…), “U+R” (update & > reconcile?) and “R” (reconcile). However, I do not know of any way to > correct the classification of a transaction while it is still in this window > (which seems like the opportunity to train the classifier). I end up > correcting the transactions after they have posted in the account registers, > and GnuCash does not seem to be learning from this. > > If you (or anyone) knows what I should be doing, please let me know - thanks! > > Cliff > > On Jun 15, 2020, at 3:25 AM, Gio Bacareza <gbacar...@gmail.com> wrote: > > Hi Cliff, > > I also want to first say thanks to all the developers. I am a user and a fan. > 2nd, awesome to meet you Cliff. I love your products too. > > I was also looking into an automated way of classifying transactions. I did a > quick automl using google and have achieved 90%+ accuracy. > > However, I wasn't able to use it because I think gnucash does its own and > requires you to classify them upon import. After a few imports it also > "learns." So subsequent imports will have these transfer accounts > autopopulated. > > I hope it helps. > > gio > > > > On Mon, Jun 15, 2020 at 5:11 PM Cliff Kushler > <ckush...@gmail.com<mailto:ckush...@gmail.com>> wrote: > Hello - > > I am a long-time user and fan of GnuCash, and start by offering my gratitude > to everyone who has contributed to it. For what it’s worth, I am also a > co-inventor of T9 and Swype, so perhaps I have contributed something to make > your life easier too. > > My question is whether there is a tool, technique or strategy to improve the > accuracy with which GnuCash assigns imported transactions to destination > accounts. I import large quantities of credit card transactions, and while > many of them are assigned to the correct accounts, a very large number of > repeat transactions (i.e. from the same vendor) are either repeatedly > assigned to the same wrong category or simply left as “Imbalance-USD”. I have > not been able to identify a way to correct them while they are still > displayed in the initial import window, although I suspect that would be the > time to correct them. I go through and manually correct them in the register > after importing them, but this does not seem to help future classification > accuracy. > > Any help would be much appreciated! I probably should have asked this years > ago… > > Joy & Blessings, > > Cliff Kushler _______________________________________________ gnucash-user mailing list gnucash-user@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.
