Dean Arnold wrote:
1) How many joins per stmt allowed ?For the moment, joins won't be changing from what they currently are - two tables with explicit joins (including outer joins) or unlimited tables with implicit joins (but those require an equijoin condition in the where clause for each table).
2) outer join support ?
Interesting idea. At the moment I am not doing much to optimize.3) re: performance: have you considered collecting COUNT(*) from each datasource
I'm not sure what the benefit of that would be over keeping it in SQL::Statement, please clue me in if I'm missing something.4) I concur that SQL extensions are better than DBI; have you considered a DBIx subclass implementation, ala DBIx::Chart ?
The only subqueries I'm contemplating so far are CREATE INTO x FROM y ... and SELECT INTO x FROM y, basically a way to dump from one rdbms to another rather than the kind of fine tuning you are talking about, but maybe in the future.5) Distributed INSERT ... SELECT, UPDATE..FROM, or DELETE ... WHERE x IN (SELECT...) supported ? (possible transactional issues, I suppose...)
6) Any minimum i/f requirements from the underlying DBD's ?
I don't think so, it will just essentially create a $dbh using the DBD and keep that stored in a lookup table.
-- Jeff