OFX example

[Wojciech Piekutowski]

On 12 August 2010 22:43, David Reiser <dbreiser at earthlink.net> wrote:
> On Aug 12, 2010, at 10:45 AM, Wojciech Piekutowski wrote:
>> Should I split BANKID and ACCTID like this or is better to stick the
>> whole account number into ACCTID and leave BANKID? Or maybe use the
>> bank name as BANKID?
>
> <BANKID> is normally the Routing/Transit Number. As long as it is unique among possible files you'll be importing, that's sufficient. I would leave BANKID and ACCTID separate.
>

Thanks, I'll try your way.

>> Is FITID used at all by gnucash, for example to match internal
>> transactions between two client's accounts?
>
> FITID is used internally to match duplicates. For a valid ofx file FITID _must_ be unique forever within a given account. Gnucash stores the FITID in the data schema and compares incoming transaction FITIDs to existing FITIDs in that account. This duplicate tracking makes ofx imports a lot more convenient if your bank always sends you 30 days of history when you download, but you actually download once a week. The FITIDs make it trivial for gnucash to automatically reject duplicate transactions.
>
> The "always unique" and "always the same for a specific transaction" makes it hard to generate FITIDs for transactions from an arbitrary bank download that doesn't contain some kind of unique identifier.

Let's say I have 2 accounts in this bank and I've transfered money
from the first to the second. Does GNUCash uses FITID to detect
internal money transfer?