Working on MT940 importer

[Christian Stimming]

Martin Preuss schrieb:
> On Friday 12 September 2003 11:50, Best, Jan-Pascal van wrote:
> 
>>Yes, I've read those. I looked at the OpenHBCI C API at
>>http://openhbci.sourceforge.net/doc/api/group__HBCI__APIg.html
>>but didn't find anything in there to parse MT940 files directly,
>>so I kind of gave up on that.
> 
> Well, the current implementation of the SWIFT parser expects a string which 
> contains the MT940. You could read a whole file into a string and pass it on 
> to the method SWIFTparser::readMT940().
> 
> However, in newer versions the parser is implemented as a plugin to OpenHBCI 
> (which we could make accessable via the OpenHBCI-API).

Sorry, Jan-Pascal, for not pointing you directly to the 
HBCI::SWIFTparser. Yes, that class has the appropriate methods.

Note about the "newer version" mentioned by Martin: His restructuring of 
OpenHBCI and its parser is not yet publicly available. He said (on 
openhbci-general) that it will probably take another month until he got 
the new structure enough in shape for committing it to openhbci CVS.

For now you can go ahead with the current OpenHBCI CVS (identical to 
0.9.13), which offers the SWIFTParser::readMT940 method.

>>It doesn't seem like there is a C wrapper
>>function available, at least not in core/swiftparser.h. Would it
>>be possible to create one?
> 
> Yes, that should not be too hard to perform.

And in fact this is quite easy. I think I will add C wrappers later today.

>>Some more things: is the booking code (you call it "transaction key")
>>interpreted somewhere? Is there a possibility to allow unstructured
>>
>>:86: fields (my bank has them)? Maybe the bank-specific parsing should
> 
> That code is not interpreted by the parser (I can't tell whether Christian's 
> GNUCash-HBCI-code makes use of it, though).

I don't think it is being used in gnucash, too.

Christian