OFX skipping transactions

[David Reiser]

You have malformed OFX files. In your .ofx files, you have many transactions with blank <FITID> fields. Those are guaranteed to be a problem.

The OFX spec requires that each transaction contain a unique <FITID>. (Unique forever for a given account, not unique in the universe of all accounts at all institutions.)

One of the blank ID transactions from the first import takes possession of the blank ID as the unique value in GnuCash's database. Any subsequent import containing blank <FITID>'s will be analyzed as containing duplicate transactions based on the blank ID as the unique ID (you have one blank ID transaction, you shouldn't ever need to import another with the same ID...).

The fact that gnucash can import the first month correctly with numerous blank ID's suggests there's an initialization issue. Or else, gnucash is assuming that no bank sends duplicates in a single download, and doesn't start checking against the current file's IDs until after the batch is complete.

Dave
On Dec 11, 2010, at 5:52 PM, Tim Richardson wrote:

> I'm trying OFX imports for the first time, since my bank now supports this format. Previously I was using QIF. The main attraction is the bayesian matching.
> It's 2.2.9 on Debian. libofx is v 0.9.0
> 
> However, many transactions are being skipped. They do not appear at all in the transaction matching screen.
> I see the missing transactions in the log file: they get a "D" in whatever the first column of the log file. This happens for a second (or later) import.
> 
> Here is a way I can reproduce the problem. Bayesian is "on", but turning it off makes no difference.
> 
> 1) Start a new gnucash set of books with a bank account.
> 2) Export two months of bank transactions in two separate files
> 3) import the first month: all transactions are entered. I let them post to unmatched.
> 4) import the second month. Many transactions are missing from the transaction matching screen.
> 
> but now:
> 5) create another new set of  books.
> 6) this time, import the second month first. All transactions appear. Import them.
> 7) next, import the first month: this time, it has lots of missing transactions.
> 
> Attached:
> 1) the log file after step 3 (test1_importApril.log)
> 2) the file file after step 4. The transactions with "D" at the start of the log file were not presented to me in the import matcher, and not imported (test1_importMay.log)
> 3) the two ofx files
> 4) the file after step 6. test2_importMayFirst.log
> 5) the file after step 7. test2_importArprileSecond.log (sorry about Arprile!) which shows transactions not importing
> 
> I guess duplicate transactions are being detected for some reason.
> 
> regards
> Tim
> 
> 
> 
> <test1_importApril.log><test1_importMay.log><201004_mainAcct_test.ofx><201005_mainAcct_test.ofx><test2_importMayFirst.log><test2_importAprileSecond.log>_______________________________________________
> gnucash-user mailing list
> gnucash-user at gnucash.org
> https://lists.gnucash.org/mailman/listinfo/gnucash-user
> -----
> Please remember to CC this list on all your replies.
> You can do this by using Reply-To-List or Reply-All.

--
David Reiser
dbreiser at earthlink.net