Re: loading only few rows from huge table
От | Bruno Wolff III |
---|---|
Тема | Re: loading only few rows from huge table |
Дата | |
Msg-id | 20041224174138.GB8243@wolff.to обсуждение исходный текст |
Ответ на | loading only few rows from huge table (Mark <sendmailtomark@yahoo.com>) |
Список | pgsql-general |
On Wed, Dec 22, 2004 at 11:31:59 -0800, Mark <sendmailtomark@yahoo.com> wrote: > Hi, > I have a table with 100K rows. One of columns is a timestamp and > indicates when this row inserted. > > What will the the best way of getting 10 latest rows from that table > and introducing partial data retrieval (rows 50-60, 100- 120, etc) If there is an index on the timestamp column you can efficiently get the 10 rows with the latest timestamps by using ORDER BY and LIMIT. Getting rows out of the middle of the table by row number efficiently is going to be harder. You can use OFFSET to get groups of records starting from number other than the latest. As the offset gets bigger this will run slower. If the timestamps are unique and you always continue from where you left off and you can save the last timestamp you retrieved, then you can use a WHERE clause to get the offset efficiently.
В списке pgsql-general по дате отправления: