Subject | RE: [firebird-support] Re: Unicode size |
---|---|
Author | Chad Z. Hower |
Post date | 2004-12-15T19:50:30Z |
:: I think Helen answered your question pretty well, I just had
:: a thought for you though. I agree with Helen that large
:: fields are generally not good candidates for indices.
Yes,but often necessary.
:: this is the case, you could store say an MD5 or SHA hash of
:: the email address field in a separate indexed field. You may
:: want to find an even simpler system, even CRC might be good
Actually that's what we used to do in the old system. I was my hopes that FB
would do this as part of its indexing instead of making us do this on each
field.. Ie index on a CRC which it looks up first, then comppares full text
of any matcing CRS matches.
I guess it does not? Maybe a good thinig to add in the future? I know other
DB's offer this. CRC 32 works well, but now it will impact all of our
lookups. We have a separate business layer so chagnes are isolated, but its
still a big change that I cannot perform right now.
:: a thought for you though. I agree with Helen that large
:: fields are generally not good candidates for indices.
Yes,but often necessary.
:: this is the case, you could store say an MD5 or SHA hash of
:: the email address field in a separate indexed field. You may
:: want to find an even simpler system, even CRC might be good
Actually that's what we used to do in the old system. I was my hopes that FB
would do this as part of its indexing instead of making us do this on each
field.. Ie index on a CRC which it looks up first, then comppares full text
of any matcing CRS matches.
I guess it does not? Maybe a good thinig to add in the future? I know other
DB's offer this. CRC 32 works well, but now it will impact all of our
lookups. We have a separate business layer so chagnes are isolated, but its
still a big change that I cannot perform right now.