What kind of text values are you searching for? Is it names, addresses or just free flowing prose?

Do you search using "contains" or "like"?

I prefer "like" when the string starts with something meaningful like "jones, tom" but
would be worthless if the first word can be something like "the old brown fox", in which case
you're stuck with "contains". I've found "like" to be way quick.


bill

Jim Limburg wrote:

G-Day all

Which situation by you all's experience would be faster. I have a table
which had about 30 columns, most of which are small - integers,1-8 size
in text and so on. One of the columns is text 100. This table has about
80,000 rows in it. The text 100 column is indexed, but searching is somewhat
slow. The table has a PK - integer. Would it be faster to project a
temp table to a user who searches if frequently, of just the PK and the
text 100 column. The user who does this searching wouldn't need it updated
each time. I could put a refresh data button on to reload the temp table,
so the user could do that a few times a day.


Or is it just faster to search on the table that has the 30 columns. I
know indexing is suppose to "put things in ram" so to speak, or at least
that's how I've figured it, so I'm not sure at this point what would be
best. I'm curious to find out some of you guru's findings in areas like
this.

Jim Limburg

.




Reply via email to