I need to create fulltext index on few columns of table (using SQL Server 2005), which length is bigger than row size (8KB) (to have one – per record unique key). Should I simply ignore row length, or is any solution for this? Situation is this:
To get (from fulltext index) row id and rank related to whole record (all text columns of record, article for example), I need to create table with all text columns in this table. But I have several big columns, which in summation exceeds 8KB row size (defined by sql server).
Current state: all texts for all languages are in one table:
idParent | langId | textType | content
-------------------------------------------------
12 | en | title | 'Title of article'
New state: to get right rank from fulltext index:
id | title | contentA | contentB | contentC | contentD | contendE | content F ...
----------------------------------------------------------------------------------
12 | 'Title' | ...
If I want to benefit of language properties of fulltext index, do I need to store data in different languages into other table/columns (to be able to set language on column)? Or is other solution to have data in one table with lang identifier (column containing lang id)?
Thank you for answer.
I did not understand your first question, can you be more clear?
Regarding the second, yes, if you want your full text to work on different languages you need to have on column for each language on your catalog. The full text is highly dependent on the collation (different languages = different collations)