On 8 November 2010 06:16, shaiju.ck <shaiju...@gmail.com> wrote: > Hi, I have a table employee with 33 columns. The table have 200 records > now. Select * from employee takes 15 seconds to fetch the data!!! Which > seems to be very slow. But when I say select id,name from empoyee it > executes in 30ms. Same pefromance if I say select count(*) from emloyee. Why > the query is slow if I included all the columns in the table. As per my > understanding , number of columns should not be having a major impact on the > query performance. I have increased the shared_buffres to 1024MB, but no > improvement. I have noticed that the query "show shared_buffers" always show > 8MB.Why is this? Does it mean that changing the shared_buffers in config > file have no impact? Can anybody help? Shaiju >
Could you run an EXPLAIN ANALYZE on the query? And what do the columns contain? For instance, if you have 10 columns each returning massive XML documents, each hundreds of megs, the bottleneck would be I/O bandwidth. -- Thom Brown Twitter: @darkixion IRC (freenode): dark_ixion Registered Linux user: #516935
