MySQL Forums
Forum List  »  NDB clusters

Re: Joins - done in NDB or MySQL?
Posted by: Gregor Melhorn
Date: January 17, 2005 05:11PM

Hello Harrison,

thanks for your reply, which has given good insight to me.

Seams like I didn't express enough what the queries do, I really don't want that huge result sets.

100 000 and 24 000 000 are the actual row counts of all data in the tables. The 24 mio rows are a zip distance lookup table, containing for each zip about 2000 other zips in a distance of up to 250 km including and ordered by distance. With MyIsam the table has a size of around 300 MB and can easily be used to look up data in other tables ordered by distance to a certain zip code without doing a full table scan (which I would have to do if I did calculations based on geodata for every request).

As this is static data, it might even be an idea to put it on the mysqld clients and take it out of the cluster, as the data won't change anyway.

My queries return around 300 rows. I was only concerned with how the pre-sorting etc is done since if there would be unneeded data pulled to the mysqld client, in this case this would be a problem.

Again thanks for your help.

Best Regards
Gregor

Options: ReplyQuote


Subject
Views
Written By
Posted
3007
January 16, 2005 06:32PM
2078
January 17, 2005 09:05AM
Re: Joins - done in NDB or MySQL?
2163
January 17, 2005 05:11PM


Sorry, you can't reply to this topic. It has been closed.

Content reproduced on this site is the property of the respective copyright holders. It is not reviewed in advance by Oracle and does not necessarily represent the opinion of Oracle or any other party.