The FILTER parameter

Use the FILTER index parameter to specify that data in a column should be filtered before it is indexed.

Important: Filtering does not modify documents that are stored in the table column; documents remain stored with proprietary formatting text.

Configure your database server for filtering, logging, and tracing before you create a filtered index. Read Configure your database server for filtering for complete details.

After you have configured your database server, you can use the FILTER index parameter in a CREATE INDEX statement to filter proprietary information from documents before they are indexed.

The following statement creates an etx index on the abstract column of the my_table table and specifies that the documents in the abstract column should be filtered before they are added to the index:

CREATE INDEX abstract_index ON my_table (abstract etx_clob_ops)
    USING etx (FILTER = 'STOP_ON_ERROR');

You can enable filtering on columns of all seven data types supported by the OneDB® Excalibur Text Search DataBlade® module: CHAR, VARCHAR, BLOB, CLOB, LVARCHAR, IfxDocDesc, and IfxMRData.

You can set the FILTER index parameter to one of the three values described in the following table.


Value	Description
NONE	Documents are not filtered before they are added to the etx index. This means that a Microsoft™ Word document, for example, is added to the index with all its formatting information. Setting the FILTER index parameter to NONE is the same as not specifying the FILTER index parameter at all.
STOP_ON_ERROR	If an error occurs during filtering, the client program that executed the statement gets a message identifying the row that caused the error, the same error is logged to the trace log (if tracing has been enabled), and the statement is stopped. See Handling filter errors for information about deciding what to do if a filter error occurred.
CONTINUE_ON_ERROR	If an error occurs during filtering, a message identifying the row that caused the error is logged to the trace log, the unfiltered document with all its formatting information is inserted into the index, and the statement continues executing. See Handling filter errors for information about deciding what to do if a filter error occurs. See Enable tracing for information about how to set up tracing.

After you have created an etx index for a table that contains data, data that you add when you insert new rows or update rows in the table is automatically filtered.

You can also filter documents in a SELECT statement by using the etx_Filter() routine. Refer to The etx_Filter() routine for more information about this routine.

For a list of file formats you can filter with the OneDB Excalibur Text Search DataBlade Module, see ids_excal_145.html#ids_excal_145.