Ahhh OK then I didn't understood a bit OK then one option is to do in two steps 1. get both start_time and max(start_time) over (partition by channel) 2. filter out only those where start_time = max(start_time) over (partition by channel) Something like: SELECT * FROM ( SELECT other_cols, start_time, max(start_time) over (partition by channel) max_start_time FROM blah ) WHERE start_time = max_start_time Gints Plivna http://www.gplivna.eu 2007/7/25, Remigiusz Sokolowski <rems@xxxxxxxx>:
Gints Plivna wrote: > A simple max(start_time) over (partition by channel) won't work? > May be I was too laconic I have thought about getting only value of rows, of which one particular value is max for the whole group - so I am not interested in rows which have this particular value less than max. "A simple max(start_time) over (partition by channel)" gives me all rows while I need only those which start_time=max(start_time) over (partition by channel). Solutions I have provided gives me exactly what I need, however I try to find more of them, then compare them and choose one... :-) Regards Remigiusz -- --------------------------------------------------------------------- Remigiusz Sokolowski <rems@xxxxxxxx> WP/PTI/DIP/ZAB (+04858) 52 15 770 MySQL v04.x,05.x; Oracle v10.x Zastrzezenie: Niniejsza wiadomosc stanowi jedynie wyraz prywatnych pogladow autora i nie jest w zadnym wypadku zwiazana ze stanowiskiem przedsiebiorstwa Wirtualna Polska S.A. --------------------------------------------------------------------- WIRTUALNA POLSKA SA, ul. Traugutta 115c, 80-226 Gdansk; NIP: 957-07-51-216; Sad Rejonowy Gdansk-Polnoc KRS 0000068548, kapital zakladowy 62.880.024 zlotych (w calosci wplacony)
-- //www.freelists.org/webpage/oracle-l