Thanks for your help. This way seems to be better. But it is still not enough I think.
-- takes about 9 seconds SELECT COUNT(ID) FROM TEST -- takes about 5 seconds SELECT COUNT(*) FROM TEST --- "Williams, Ken" <[EMAIL PROTECTED]> wrote: > > > > -----Original Message----- > > From: Ali Sadik Kumlali [mailto:[EMAIL PROTECTED] > > Sent: Monday, March 29, 2004 3:09 PM > > To: [EMAIL PROTECTED] > > Subject: RE: [sqlite] Is using max(rowid) instead of count(*) safe? > > > > > > Thanks for taking your time. I'm using P3 600Mhz, 512MB, Win2K Pro > for > > the test. Here is what I've done: > > > > CREATE TABLE TEST( > > ID NUMBER(12) PRIMARY KEY, > > NAME VARCHAR(19) > > ) > > > > -- Transfers 3.5 million rows > > COPY OR IGNORE TEST FROM 'C:\\DATA.CSV' USING DELIMITERS ';' > > > > -- takes about 9 seconds > > SELECT COUNT(ID) FROM TEST > > Try again with COUNT(*) instead of COUNT(ID) if you just want the > number of > rows. The latter needs to figure out which rows have ID NOT NULL, > but the > latter just counts all rows. > > I'm surprised that COUNT(ID) would take 9 seconds, but you could also > try > creating the table with ID's type as "INTEGER PRIMARY KEY NOT NULL" > so that > you're using the BTree table index directly. __________________________________ Do you Yahoo!? Yahoo! Finance Tax Center - File online. File on time. http://taxes.yahoo.com/filing.html --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]