| From: | Daniel Ceregatti <vi(at)sh(dot)nu> |
|---|---|
| To: | pgsql-performance(at)postgresql(dot)org |
| Subject: | What is the best way to do attribute/values? |
| Date: | 2004-08-19 18:03:04 |
| Message-ID: | 4124EB58.9040102@sh.nu |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-performance |
Hi list,
I have a database with 1M "people" in it. Each person has about 20
attributes, such as height, weight, eye color, etc. I need to be able to
search for people based on these attributes. A search can be conducted
on one attribute, all attributes, or any number in between. How would
_you_ do this?
I have already attempted to answer this. My attempts are detailed here:
This is the email I was originally going to send to this list. Since
it's so large, I decided to link to it instead. If you feel that it
belongs in a post to the list, let me know, and I'll post again.
I've discussed these attempts with people in #postgresql on
irc.freenode.net. Agliodbs (I presume you know who this is) was very
helpful, but in end was at a loss. I find myself in the same postition
at this time. He suggested I contact this list.
My ultimate goal is performance. This _must_ be fast. And by fast, I
mean, < 1 second, for every permutation of the number of attributes
searched for. Flexibility would be a bonus, but at this point I'll
settle for something that's harder to maintain if I can get the speed
gain I need.
Thanks,
Daniel Ceregatti
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Richard Huxton | 2004-08-19 18:09:55 | Re: using an index worst performances |
| Previous Message | Gaetano Mendola | 2004-08-19 17:26:01 | using an index worst performances |