Hi,
I'm looking for a solution how to import splitted transactions into existing KMyMoney file/db. I want to import transactions from 3rd party DB which contains both single (simple) transactions and also splitted transaction. In this case I can choose any intermediate format used to import data into KMyMoney.

When exporting to CSV details about splitted transaction are appended to end of row which contains splitted transaction (splitCategory, splitMemo splitAmount).

Is there any way to import split-transactions?
If so:

• Which format supports it? CSV? OFX? QIF?
• How should be splitted transaction defined in that file?
  
  • In case of CSV - can be used same format as in exported CSV? I dont think so, because in import wizard there is no option to assign columns for split Category/Memo/Ammount.

If not:

• What do you recommend?

AFAIR, QIF is the only format that supports the import of split transactions. As stated on the QIF Definition page:

Note: Repeat the S, E, and $ lines as many times as needed for additional items in a split.

What you can do to see how such transactions look like in QIF is creating one in a 'demo' account in KMyMoney and then export it to QIF.

CSV using additional columns for splits is not supported. Not sure if OFX supports splits at all.

p.s. nice nick name

Thank you! QIF looks as the right choice, specially because last paragraph "Special KMyMoney QIF extensions" applies to my application. So I can overcome one of drawbacks this format has.

I have two more questions about this special '#' field:

1. Do you know where content of # field is stored inside KMyMoney database? I was thinking to insert my unique transaction identifier inside `kmmSplits.checkNumber`.
2. Is there any limitation for text length in # field?

Of course I know (ok, can find out) where KMyMoney stores that information. It is kept in an internal field and not accessible to the user. Usually, this information is provided by your institution when you download transactions online from your bank's server (using OFX or HBCI).

Yes, I suppose that transaction unique ID will not be accessible to the user through GUI. I want to use it to detect which transactions from 3rd party SW was already imported to the KMyMoney and which transactions should be exported to QIF for next import. I'm using MySQL as KMyMoney storage so I want to look into DB and make QIF export fully automated. Could you help me with identification of this column? Afterwards I can study KMM sources and find the QIF '#' field limitations.

Finally I realized that transaction identifier from KMyMoney custom tag '#' in QIF is stored in db colum 'kmmSplits.bankId'. Data type is MEDIUMTEXT so there is almost no limitation

Automated export to QIF and import to KMM works now as expected, thank you!