Thanks for the answer. If I go that way, though, I am afraid I will have very long query times. And I have no idea how to handle large databases of that kind without losing performance. If I store the pattern - is there a way to still perform the query in the database? Or do I have to do that in my software? (Matching the resultset to the pattern) Thanks Matt --- Ursprüngliche Nachricht --- Datum: 16.11.2004 16:42 Von: Richard Huxton <dev@xxxxxxxxxxxx> An: matthias@xxxxxxxxxx Betreff: Re: How to deal with almost recurring data? > matthias@xxxxxxxxxx wrote: > > I am creating a database which is supposed to contain many data entries > > (events) that differ only in the date they occur. > > > > So let's say event 1 occurs every Monday, Tuesday and Sunday between > > January 1st and May 30th 2005. > > > > How do I store and manage such data in a meaningful way? > > > > The simple idea would be to store the event itself in one table and have > > another table containing all the dates (all Mondays, Tuesdays and Sundays > > between 2005-01-01 and 2005-05-30) plus a foreign key to event_ID => > > (date, event_id). > > > > The problem is that we are dealing with several tenthousand events, > > resulting in several million single dates if I stored it in the described > > manner. > > This seems like a reasonable approach. Don't forget that PostgreSQL is > designed to store large amounts of data. > > Of course, if there is some pattern to the dates then you might be > better off storing the pattern rather than the individual dates. > > -- > Richard Huxton > Archonet Ltd > > ---------------------------(end of broadcast)--------------------------- > TIP 6: Have you searched our list archives? > > http://archives.postgresql.org > ---------------------------(end of broadcast)--------------------------- TIP 8: explain analyze is your friend
