The sort order specifies the collating sequence used to order the data in the result set. The default sort order is set by the sort_order configuration parameter (for more information, see “Setting the Default Sort Order”). Case insensitive sort order is supported in the Enhanced version.
Use the sort_by pseudo column to return a result set with a sort order other than the default. With the Enhanced Full-Text Search engine, you can specify up to 16 sort specifications in the sort_by pseudo column.
Table 5-2 lists the values for the sort_by pseudo column.
Value |
Description |
|---|---|
fts_score desc |
Returns a result set sorted by score in descending order. |
fts_score asc |
Returns a result set sorted by score in ascending order. |
fts_timestamp desc |
Returns a result set sorted by a timestamp in descending order. |
fts_timestamp asc |
Returns a result set sorted by a timestamp in ascending order. |
column_name desc |
Returns a result set sorted according to the descending order of a column. column_name is the name of the source table’s column. |
column_name asc |
Returns a result set sorted according to the ascending order of a column. column_name is the name of the source table’s column. |
fts_cluster asc |
Returns a clustered result set. Only available with the Enhanced Full-Text Search engine. (See “Using Pseudo Columns to Request Clustered Result Sets” for more information.) |
Before you can sort by specific columns, you must modify
the style.vgw and style.ufl files
(see “Setting Up a Column to Use As a Sort Specification”).
For example, the following query sorts the documents by timestamp in ascending order:
select t1.score, t2.copy from i_blurbs t1, blurbs t2 where t1.id=t2.id and t1.score > 90 and t1.index_any = "<accrue>(raconteur, Paris)" and t1.sort_by = “fts_timestamp asc”
