Hi there,
after tracing the sql generated I found out, that hslqdb behaves
completely different depending on where the join-clause is.
The statement
SELECT DISTINCT document.ID, document.ARCHIVE, document.ARCHIVEID, document.NAME,
document.DOCCONFIGID FROM document, attribute
WHERE UPPER(attribute.LANGUAGE)=UPPER('en') AND
UPPER(attribute.VALUESEARCH) = UPPER('ABBIndustrie') AND
attribute.CONFIGID=17 AND
document.ID=attribute.DOCID
takes 50 seconds.
The statement
SELECT DISTINCT document.ID, document.ARCHIVE, document.ARCHIVEID, document.NAME,
document.DOCCONFIGID FROM document, attribute
WHERE document.ID=attribute.DOCID AND UPPER(attribute.VALUESEARCH) =
UPPER('ABBIndustrie')
AND UPPER(attribute.LANGUAGE) = UPPER('en')
and attribute.CONFIGID=17
takes muss less than one second.
The difference is that in the second one the join clause is at
the beginning of the where clause.
Well, before pathing torque to change the order I will address myself
to the hsqldb-team,
Thanks for your help!
Patrick
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
