query execution question - Mailing list pgsql-hackers

From amit sehas
Subject query execution question
Date
Msg-id 872907.86083.qm@web32203.mail.mud.yahoo.com
Whole thread Raw
Responses Re: query execution question  (Nicolas Barbier <nicolas.barbier@gmail.com>)
Re: query execution question  ("Kevin Grittner" <Kevin.Grittner@wicourts.gov>)
List pgsql-hackers
We are making some performance measurements, we are trying to determine query execution behavior. Lets say we have 4
tablesT1, T2, T3 and T4
 

and the query has the form:

select * from T1, T2, T3, T4 where (T1.a = T2.b   and  T2.c = T3.dT3.e = T4.f)  

where a,b,c,d,e,f are properties of the respective tables.

Lets say that the cost based optimizer determines that the order of the joins should be   T1.a=T2.b followed by T2.c =
T3.dfollowed by T3.e = T4.f
 

the question we have is during query execution are the joins evaluated completely one by one in that order, or the
firstjoin is evaluated completely and generates an intermediate table which is then utilized to perform the next
join....thismeans that for such a query we will need space for all the intermediate tables, which if they are very
largetables as they are in our case can significantly alter the cost of the operations...
 

thanks

-Ashish


     


pgsql-hackers by date:

Previous
From: Robert Haas
Date:
Subject: Re: SSI patch version 14
Next
From: Chris Browne
Date:
Subject: Range Types - efficiency