Full text - search a table using another table?

Apologies if this is a really basic thing - I read a couple books'
chapters on full text and didn't see it.

I have table A which has the event log for the server
Table B is a blacklist - phrases to ignore. "The server has been on X
days", that type of thing.

I'd like to either search table A for phrases in table B (and then
delete them), or search A where not exists phrases from table B, and
insert that into a new table.

Is this doable? If so, how? Many thanks.

On Jan 15, 12:48 pm, "Hilary Cotter" wrote:
> I am not sure what you want to do here.
>
> Is this what you want - search on a prhase, if the phrase is found in table
> b, its blacklisted so don't search on table a.
>
> If it is not found in table b then search on table a?

Apologies for not being clear.

I want to pull all records in table A that aren't in table B. But
it's not an exact match so I can't just do a NOT EXISTS. I figured
that going the Full Text route would allow me to do a search.
Table A:
The ABC service entered the paused state.
The ABC service entered the running state.
Hi Mom
The WinHTTP Web Proxy Auto-Discovery Service service was successfully
sent a start control.
The WinHTTP Web Proxy Auto-Discovery Service service entered the
running state.
SQL Trace ID 2 was started by login "NT AUTHORITY\SYSTEM".
SQL Trace stopped. Trace ID = '2'. Login Name = 'NT AUTHORITY
\SYSTEM'.
The ABC service encountered an error.

Table B:
The ABC service entered the
Web Proxy Auto-Discovery Service
SQL Trace

Results desired:
Hi Mom
The ABC service encountered an error.

Practically, what I'm trying to do is query all the event logs in our
environment, and add filters to a table as we find things we don't
need to see in subsequent runs. I don't want to do an exact match
since there are events with unique strings ("The server has been
running X days", for instance), and I'd rather everything be handled
via this blacklist table rather than manually coding exclusions.

One alternate idea was to build a varchar(max) with a series of OR
statements all chunked together, but that's going to get cumbersome
and slow very fast. Another idea is to build an index-of-strings,
chopping every 3 words and trying to exclude by 3-word phrases (though
again, for some I'd have to manually code exclusions)


Finally, this is on 2005, though I've read from you (October 28,
social.msdn.microsoft.com) that in SQL 2008 you can't force FTI to do
an exact match, it will always do a fuzzy search, which might be a
concern when we upgrade.