Saving with new mysql 5.7 triggers in db

[John Ralls]

> On May 15, 2017, at 8:49 AM, Derek Atkins <warlord at mit.edu> wrote:
> 
> Geert Janssens <geert.gnucash at kobaltwit.be <mailto:geert.gnucash at kobaltwit.be>> writes:
> 
>>> Is it possible that GnuCash just isn't being happy with the "modified"
>>> tables?
>>> 
>> Possibly and that's what John also suggests. I don't know the exact details of 
>> how the sql backend interacts with the tables. I just know there are ways in 
>> general that you can have extra columns in an sql table than there are columns 
>> used in an insert/update query. When columns are omitted they are normally set 
>> to their default value on insert or ignored on an update. I used that idea to 
>> suggest my alternative trigger.
>> 
>> On the other hand if the sql backend for some reason sets restrictions on the 
>> available columns this may be an issue. Only a real test can tell but I'm not 
>> interested enough right now to spend the effort. On the other hand if someone 
>> else is, I would appreciate to hear the result.
>> 
>> And John's other remark was even more to the point. Why not use the year(), 
>> month() and day() function directly in the queries instead of duplicating the 
>> info in the table ?
> 
> I haven't looked at the code but inserts can be done implicitly or
> explicitly.  If they are done implicitly then any change to the table
> could cause problems on future inserts.  I don't know if GnuCash uses
> implicit or explicit insertions.

If "implicit" means "whole record with fields in order and not tagged with field names" that's what GnuCash does.

Regards,
John Ralls