Hi,

I have a table with more then 500000 records.
I'm trying to search for a part of word in a description-column (char50),
normally I take the like statement. But sometimes it is too slow.

Does anyone have a solution for that ?
Using full text search is not an option.

Thx
Geert

Did you mean varchar(50)? If it is really char(50), are they really mostly
50 characters? If not, you may see some minor improvements if you switch to
varchar.

Store the data better? If individual words are important then maybe you
could consider a relational table with the individual words where you can
search with equality instead of wildcards. This is similar to the "tagging"
concept where you can tag a row with 1-n tag words, and they are stored
separately so it is more efficient to search for them (using an index) as
opposed to parsing a delimited list.

Why? This is exactly what it's for. And since FTS is even available in SQL
Express w/advanced services, can you explain why it is not an option?

Did you mean varchar(50)? If it is really char(50), are they really mostly
50 characters? If not, you may see some minor improvements if you switch to
varchar.

Store the data better? If individual words are important then maybe you
could consider a relational table with the individual words where you can
search with equality instead of wildcards. This is similar to the "tagging"
concept where you can tag a row with 1-n tag words, and they are stored
separately so it is more efficient to search for them (using an index) as
opposed to parsing a delimited list.

Why? This is exactly what it's for. And since FTS is even available in SQL
Express w/advanced services, can you explain why it is not an option?

Did you mean varchar(50)? If it is really char(50), are they really mostly
50 characters? If not, you may see some minor improvements if you switch to
varchar.

Store the data better? If individual words are important then maybe you
could consider a relational table with the individual words where you can
search with equality instead of wildcards. This is similar to the "tagging"
concept where you can tag a row with 1-n tag words, and they are stored
separately so it is more efficient to search for them (using an index) as
opposed to parsing a delimited list.

Why? This is exactly what it's for. And since FTS is even available in SQL
Express w/advanced services, can you explain why it is not an option?

Did you mean varchar(50)? If it is really char(50), are they really mostly
50 characters? If not, you may see some minor improvements if you switch to
varchar.

Store the data better? If individual words are important then maybe you
could consider a relational table with the individual words where you can
search with equality instead of wildcards. This is similar to the "tagging"
concept where you can tag a row with 1-n tag words, and they are stored
separately so it is more efficient to search for them (using an index) as
opposed to parsing a delimited list.

Why? This is exactly what it's for. And since FTS is even available in SQL
Express w/advanced services, can you explain why it is not an option?

Did you mean varchar(50)? If it is really char(50), are they really mostly
50 characters? If not, you may see some minor improvements if you switch to
varchar.

Store the data better? If individual words are important then maybe you
could consider a relational table with the individual words where you can
search with equality instead of wildcards. This is similar to the "tagging"
concept where you can tag a row with 1-n tag words, and they are stored
separately so it is more efficient to search for them (using an index) as
opposed to parsing a delimited list.

Why? This is exactly what it's for. And since FTS is even available in SQL
Express w/advanced services, can you explain why it is not an option?

Did you mean varchar(50)? If it is really char(50), are they really mostly
50 characters? If not, you may see some minor improvements if you switch to
varchar.

Store the data better? If individual words are important then maybe you
could consider a relational table with the individual words where you can
search with equality instead of wildcards. This is similar to the "tagging"
concept where you can tag a row with 1-n tag words, and they are stored
separately so it is more efficient to search for them (using an index) as
opposed to parsing a delimited list.

Why? This is exactly what it's for. And since FTS is even available in SQL
Express w/advanced services, can you explain why it is not an option?

Did you mean varchar(50)? If it is really char(50), are they really mostly
50 characters? If not, you may see some minor improvements if you switch to
varchar.

Store the data better? If individual words are important then maybe you
could consider a relational table with the individual words where you can
search with equality instead of wildcards. This is similar to the "tagging"
concept where you can tag a row with 1-n tag words, and they are stored
separately so it is more efficient to search for them (using an index) as
opposed to parsing a delimited list.

Why? This is exactly what it's for. And since FTS is even available in SQL
Express w/advanced services, can you explain why it is not an option?

Did you mean varchar(50)? If it is really char(50), are they really mostly
50 characters? If not, you may see some minor improvements if you switch to
varchar.

Store the data better? If individual words are important then maybe you
could consider a relational table with the individual words where you can
search with equality instead of wildcards. This is similar to the "tagging"
concept where you can tag a row with 1-n tag words, and they are stored
separately so it is more efficient to search for them (using an index) as
opposed to parsing a delimited list.

Why? This is exactly what it's for. And since FTS is even available in SQL
Express w/advanced services, can you explain why it is not an option?

Did you mean varchar(50)? If it is really char(50), are they really mostly
50 characters? If not, you may see some minor improvements if you switch to
varchar.

Store the data better? If individual words are important then maybe you
could consider a relational table with the individual words where you can
search with equality instead of wildcards. This is similar to the "tagging"
concept where you can tag a row with 1-n tag words, and they are stored
separately so it is more efficient to search for them (using an index) as
opposed to parsing a delimited list.

Why? This is exactly what it's for. And since FTS is even available in SQL
Express w/advanced services, can you explain why it is not an option?