Index database files (*.dbf, *.csv, *.mdb, and *.accdb) as text, without applying field attributes to the content or separating rows into separate documents.

Use dtsoTfDatabasesAsText to suppress field attributes in database files while indexing all of the text. Using this flag will make indexing and searching faster and more efficient when database field searches are not needed. 

By default, each row of a database file is indexed as a separate document, field searches can be done using the field names in the database to retrieve individual rows. For example, if a database contains a "Last_Name" field, then you could search for "Last_Name contains Smith" to find all database rows with "Smith" in the Last_Name field. 

In some cases, such as forensic or E-discovery searching, users want to search across all text without regard to field attributes, so indexing each row as a separate document and indexing using field attributes provides no benefit.