Re: [BUGS] CREATE INDEX spoils IndexScan planns

2003-10-31 Thread Nitz
Hi Rod, here is the actual production trace of the problem. This is a table of mobile network cells and code-names devided into LAC's. Two test cells to test with are: test cell id #1: 900 4900035 test cell id #2: 300 5080140 You were right, the volume of the data changes the optimizer's willingn

[BUGS] CREATE INDEX spoils IndexScan planns

2003-10-31 Thread Nitz
Hi! SOME BACKGROUND: I am having a rather annoying problem which I have no explonation to so I am calling it a bug. Currenty I am designing a rather huge database with 700+ tables and unfortunately I am quite late to discover this as my test data is imported via \i the_whole_database_definition