Shouldn't it be:
select name, day, distance
from (
select name as max_name,max(distance) as max_distance
from t
group by name
) as foo inner join t on t.distance = foo.max_distance and
t.name=foo.max_name
and this still returns two rows when someone has run the same
It seems to me that MySQL and PostgreSQL are exhibitting some dubious
guessing behaviour as to which column it refers to, or perhaps they are
making some requirement of the order by to contain a column that is in the
resultset ??? If name were only in test11 then what would MySQL and
PostgreSQL do
Nevermind, it seems I have found the cause, I was doing:
sqlthread->exec("PRAGMA synchronous = 0;",NULL,NULL);
sqlthread->exec("PRAGMA cache_size = 4000;",NULL,NULL);
but changing it to this has sorted the problem out:
sqlthread->exec("PRAGMA
I've just upgraded my program from using v2.8.16 of sqlite to using v3.2.1
of sqlite. So I've change the wrapper class that I was using to use the new
prepare() instead of compile() etc. The problem I have is that my program is
now running slower - a lot slower, as in one fiftieth of its previous
I am also having difficulty getting to grips with the (awkward) join
notation. In 3.0 how about support for Oracle type joins with (+) which
seems far simpler to me :)
-Original Message-
From: D. Richard Hipp [mailto:[EMAIL PROTECTED]
Sent: 12 April 2004 13:00
Cc: [EMAIL PROTECTED]
5 matches
Mail list logo
