Searching and the SQL Server database

In SQL Server, the collations provided by the Windows operating system are effective in providing language-appropriate searching. To work correctly, Guidewire requires that you create a SQL Server database with case-insensitive collation. Guidewire uses this collation for all character data sorting and searching by default, as well as to provide case-insensitive table and column names.

Through the linguistic search configuration, it is possible to specify a different collation for searching on columns that support linguistic searching:

  • If simple, case-insensitive searching meets your requirements, then configure collations.xml to select the same collation as the database collation.
  • If you need different search semantics, then configure the SQL Server entry in collations.xml for a primary strength search collation, which will give you accent-insensitive searching.

The semantics of linguistic searching for SQL Server are those of the Windows collation selected from the collations.xml file. The collation is based on the default language and linguistic search collation strength from language_languageCode.xml, in which secondary strength is the default. Microsoft controls the Windows collation rules, not Guidewire.

With reference to the discussion about Japanese and German search rules on Oracle, the Windows collations configured in the base configuration in collations.xml provide the following:

  • Kana-insensitivity and width-insensitivity for Japanese collations
  • Umlaut and Eszett handling in the German collations

If you are currently using SQL Server in those languages, your IT staff is mostly likely familiar with these issues.