On 05.03.2014 11:02, RSmith wrote:
On 2014/03/05 10:41, Dominique Devienne wrote:
On Tue, Mar 4, 2014 at 9:41 PM, Elefterios Stamatogiannakis
One thing that IMHO long term might improve the situation would be if
SQLite's own "native" tables would use the same Virtual Table API,//...
...//Of course, the above is a "naive" abstract reflection which ignores
the realities of VDBE, so it may sound rubbish to actual SQLite
developers. Apologies for that. --DD
I don't think it is rubbish at all, but maybe idealistic. The biggest
problem I can see from making API's pov is that you can at any time
alter, update, change the way SQLIte (or any other API) works with the
base check that the input values produce the same (or maybe
more-correct) results. Once you let the VT use the same API, any change
is a potential change to how other people's programmed interfaces need
to talk to - or get data from - the SQLite engine. This cannot simply
change on a whim, so the levels of separation remain needed.
That said, I'm all for making a more efficient VT API, but it would
probably need to be "new" functionality since I cannot see how the
existing interface could implement any of the mentioned enhancements
without breaking existing behaviour. The OP's xNextRow suggestion seems
a good idea, but opens up a whole can of what-ifs which other posters
have alluded to, but something to that effect might be worthwhile if the
efficiency bonus is significant.
The whole thread so far is based on the OP observations in mixed C/PyPy
and apsw/CPython environments (as being said already, we suffering the
noise in both cases).
To be helpful to the SQLite team, before proposing any changes, please
let someone show some well designed, pure C vtable implementation
demonstrating the possible vtable interface inefficiency.
Let's remember that all xNextRow, xNextPage optimizations are
applicable only for "select *" cases, not in the general "select f(x),
y" case.
Kind regards,
Alek
_______________________________________________
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
