Import wizard's "Smartness"

[Oliver Schinagl]

 On 10/11/10 15:00, Christopher Singley wrote:
> On Sun, Oct 10, 2010 at 8:03 PM, Oliver Schinagl <oliver at schinagl.nl> wrote:
>>  On 10/11/10 02:21, FireFly wrote:
>>>>>> So to summarize, what should I do to make gnucash
>>>> smarter
>>>>>> when importing data? I can modify the csv file
>>>> slightly (and
>>>>>> the perl script) to make minor adjustments but I
>>>> really am
>>>>>> not looking forward on working on 5 years of
>>>> bankstatements
> Can't speak for gnucash import logic, but your generated OFX looks
> malformed.  <BANKACCTTO> is just an aggegating element that consists
> of just a tag, without any text data.  It contains three child
> elements ( <BANKID>, <ACCTID>, and <ACCTTYPE>) which are used as a
> 3-tuple to uniquely identify a bank account.  This might be screwing
> up your import.
>
> This part of the spec is stable across different OFX versions; you can
> find it e.g. on pg. 184 of the OFXv2.0.3 spec:
>
> http://www.ofx.net/DownloadPage/Downloads.aspx
>
> You may also wish to look closely at Chpt. 12 of the spec, concerning
> online payments - it's possible that this is actually a better to what
> your transactions are & what data your bank are offering for download.
I'll read into that.

Ideally, an example with one 'perfectly importable' ofx would help a lot
I suppose; i'll try searching for it, but since gnucash can't export OFX
it's hard to 'imagine' it.

Thanks though!