Invoice Importer

Derek Atkins warlord at MIT.EDU
Mon Apr 2 09:10:06 EDT 2012


Reuben Cummings <reubano at gmail.com> writes:

>>>>> SELECT
>>>>> 	entries.invoice
>>>>> 
>>>>> 	FROM entries
>>>>> 	LEFT OUTER JOIN invoices ON entries.invoice = invoices.guid
>>>>> 	WHERE invoices.guid IS NULL
>>>>> 	GROUP BY entries.invoice
>>>> 
>>>> Do you really need the join here?
>>> 
>>> Yes. Since the orphan entries are identical to the non-orphan entries,
>>> the only way to filter them out is to join on the invoices table and
>>> exclude all the entries that match. That leaves entries with
>>> non-matching invoices... the orphan entries.
>> 
>> Really?  Why can't you just do:
>> 
>>  SELECT ... FROM entries where entries.invoice IS NULL?
>
> Because entries.invoices isn't null. Even the orphan entries have an invoice guid. It just that these guids are not present in the invoices table.

Oh, so the issue is that there was an Invoice that was created in RAM
but never committed.  I see.  That could never happen through the GUI,
but the importer clearly bypasses the GUI.  So the Entries have a
reference to a non-existing Invoice.   Very Interesting.

Thanks for working through this with me.

-derek

-- 
       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/    PP-ASEL-IA     N1NWH
       warlord at MIT.EDU                        PGP key available


More information about the gnucash-devel mailing list