I should have mentioned this previously: running the same query against the same data on 8.1.5 does not result in a hash aggregate plan or an out of memory error. (Note: the hardware is different but very similar -- the main difference is the 8.1.9 server (with the error) has faster disks)
On 6/14/07, Mason Hale <[EMAIL PROTECTED]> wrote:
Thanks Tom. Here's more info: What have you got work_mem set to? 40960 What's the actual number of groups > (target_page_id values)? Approximately 40 million (I'll have a more precise number when the query finishes running ). Maybe this helps? crystal=> select null_frac, n_distinct, correlation from pg_stats where tablename = 'page_page_link' and attname = 'target_page_id'; null_frac | n_distinct | correlation -----------+------------+------------- 0 | 550017 | 0.240603 (1 row) Mason