In my particular situation, count(1) or count(my_column_name) does not seem to make a dent. I understand that count(x) is useful for getting the row count of a subset of results. But since I just want the number of rows in a table, shouldn't the query just be able to get this from a value associated with this table? Is this value maintained in the sqlite code? Buzz
Hennie Peters <[EMAIL PROTECTED]> wrote: At 07:41 6-1-2004 -0800, Buzz Weetman wrote: >I'm using sqlite in an embedded system. Disk I/O is relatively slow. I >am doing a: >SELECT count(*) FROM my_table; > >This is taking too much time. I've also tried count(1), as I saw >suggested elsewhere in the mail list... though I'm not entirely sure what >this means. 1st column? I've tried a column that I have an index for... >not faster. > >Doesn't sqlite "know" the number of rows in each table without explicitly >counting them? > >Thanks for any help >Buzz The count function takes a column or a constant as an argument. I posted the count(1) some days ago. I used this in Oracle sql years ago. The idea was to make the query faster as it did not need to scan the columns in the table. groet, Hennie ================================================== Linux is like a wigwam - no gates, no windows and an apache inside. ================================================== --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------- Do you Yahoo!? New Yahoo! Photos - easier uploading and sharing
