Hi,
I am using MySQL and doing "free text searches" on keywords in a
VARCHAR(120) colum. Currently the number of records are about 1000000 and
frequently changing.
The query operating on the column is of the form "WHERE FOO LIKE %BAR%".
And I was advised to do
1. use "text search" product such as Altavista or Excite. or
2. Learn about "inverted hash trees" and other non-B-Tree based indexing
schemes and roll our own using MySQL as the core.
It's good if I could use some of the product. Does Harvest or Altavista
works for MySQL?
Making "inverted hash trees" really solves the problem? If it really worth,
then I could try. But are there any measure that how much it could make the
query fast?
Actually, our company is doing search engine for auction items such as ebay
and AuctionWatch. They are really fast with more items we have and we are
realy wondering how they are making it.
I have been wondering if it is possible to achive that much of speed by just
connecting user input from web to some database and have proper index
tuning...
Insider information welcomed!
PS,
I tried to make "wordmap" and made as a cash. There is a log of 377613
number of past searched keywords and I made a kind of cash of wordmap as
word->wordmap->maindata. This wordmap gets huge very easily. But by making
wordmap for most frequent 350 word, those could covers 10 percent of the
whole user queries. Which is helpful. The reason why we could generate
about 350 words are because of the frequenty of the change of data. We are
getting about 500 data in 10 minutes, as I remember and about the same could
be expiring. That's why we need "optmize data" frequently and do do it for
huge wordmap takes too long time.
Thanks,
________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com
______________
If you want to unsubscribe send "unsubscribe udmsearch"
to [EMAIL PROTECTED]
