[GNC] Transactions association to accounts (automatically)

David Cousens davidcousens at bigpond.com
Sun Mar 1 15:48:51 EST 2020


Stelvio,

The import matching process is used by most of the import methods including
CSV. 

The problem when you start out is that it can't initially know which
specific accounts you wish to allocate transactions to, but it accumulates
information as you import data that allows GnuCash to automatically suggest
the correct account based on past import choices. It uses either an older
mapping approach or a Bayesian probability methodology to assign the
accounts based primarily on information in the Description anda few other
fields in the import record. 

Edit-> Preferences->Online Banking will display the preference settings for
the matcher which is also used on data downloaded live from banks using
aqbanking.  AFAIK the "Use bayesian matching" is selected by default but if
it isn't make sure it is checkeded. It is probably not advisable to change
any of the other default  selections until you are more familiar with
importing data.

It also searches through existing transactions within a time window about
the import date (~6 weeks) and identifies transactions for which the
imported data might be duplicates and then tries to match a transfer account
if this is not specified explicitly specified in the import record (it can
be). 

You can manually override any of the automatic duplicate matches and account
assignments in the Import Matcher window. When GnuCash first starts out it
also constructs a mapping of external Account Names in imported data to the
internal GnuCash names which correspond to them. If an account name is
specified for the first time in import data you will get a dialog asking you
to match it to an Account in the account heirarchy you have setup in
GnuCash. Once you have done that GnuCash will remember the association.

The import training only takes place from data as it is imported and does
not use any transaction information already in Gnucash. This is important
because the import matching data is stored in the last stage of the import
process and if you don't complete an import GnuCash won't accumulate data
from that import on which these automatic processes are based. It is also
necessary to make sure any duplicates identified are actually duplicates -
the matcher will sometimes get it wrong and to make sure the correct
transfer accounts have been assigned manually before importing otherwise the
matching process does not get trained.

See https://www.gnucash.org/docs/v3/C/gnucash-help/trans-import.html for
more info.  There is updated documentation on the way on this as well but my
progress on this lately has been limited by health and other issues.

if you are using one of the v3.x GnuCash version, the CSV import is vastly
improved particularly in the setup of options for the import of data with
the ability to store and recall setups for specific data sources. If you
experiment with this you should be able to minimize or eliminate any
preprocessing in Excel before importing  data. Another possibility is to
setup and use macros in Excel or LibreOffice to preprocess your data
automatically and prepare it for import.

David Cousens



-----
David Cousens
--
Sent from: http://gnucash.1415818.n4.nabble.com/GnuCash-User-f1415819.html


More information about the gnucash-user mailing list