Client Support
cancel

Confirmation on required character set for the TRIM database.

Highlighted
Joshua Hutley
Outstanding Contributor.

Confirmation on required character set for the TRIM database.

 

Our current character set on TRIM production SQL database is SQL_Latin1_General_CP1_CI_AI.

 

Is this correct for TRIM 7.10 sql database?

 

What about a 7.34 SQL database

 

HP Documentation I have from 2010 recommends character set should be Latin1_general_CI_AI.

 

 

Thanks

2 REPLIES
EWillsey
Acclaimed Contributor.

Re: Confirmation on required character set for the TRIM database.

In a sense they are basically the same.  "SQL_Latin1_General_CP1_CI_AI" is the SQL collation with insignificant accents. "Latin1_General_CI_AI" is the Windows collation with insignificant accents.  The different comes down to the sorting and seeking rules for the varchar datatypes (think of all the string fields in TRIM). 

 

There is a performance hit when using Latin1_General_CI_AI and comparing varchar to nvarchar, but nvarchar is not used within TRIM.  Everything is a varchar.  So right off the bat there's no performance problem with the Windows collation.  IT also doesn't get impacted when using unicode characters.

 

The SQL collation, however, does not support index seeking if unicode is used.  So you have a performance hit.  You'll notice this in TRIM when unicode is enabled, SQL Collation is used, and you perform a search against a table indexed on a varchar field (this title word here).  The results may not even be noticable to your uses.  Only by using SQL query analyzer can you see the performance hit.... or in very large databases.

 

There's a lot of other stuff in the collation, but it's really boring to discuss.  Hopefully this helps a little.

 

 

Joshua Hutley
Outstanding Contributor.

Re: Confirmation on required character set for the TRIM database.

Thank you Erik

 

This is helpful. Our DBA confirmed we don't have any unicode data types and HP have confirmed this character set is supported