On 2014-03-13, 8:17 AM, big stone wrote:
Hello again,
Windowing functions are :
- not easy to workaround with create_function(),
- a truly "sql core" motor functionality.
The only use case that I have, is the small subset I described earlier :
- inside one 'select' :
. several sum/min/avg/max (different fields) ,
. all with exactly the same "(over partition by ... a series of fields)",
- no ranking, no order by inside this partition.
==> It allows to show a list of records at a certain level, with
statistical analysis done at a completely different (higher or lower) level.
Is it a feature SQLite team would like to do ?
Is there anyone else, (besides little bee), that would "like" this request
?
I know I would like to see this, a lot.
Its one of those things that, if implemented in the core, should not be any more
effort than it takes to implement aggregate functions with GROUP BY, and it
would give users a great amount of power.
I was very happy to see WITH get into the core, and windowing is similarly
something you can get a lot of power from with relatively small core effort.
If SQLite does this, I will be happy that there is yet another significant way
in which SQLite is more powerful than MySQL (but not PostgreSQL), the WITH
support being another, and subjecting data definition to transactions is another.
-- Darren Duncan
_______________________________________________
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
