Next: , Previous: Introduction, Up: Tutorial


1.2 Indexing

Here is a simple indexing command (`$' is the UNIX prompt):

     $ af -i -d mydb -C -v *.txt

The -i option indicates that we want to index files, and the list of files is expected to be at the end of the line (*.txt). The -d option specifies the database name, in this case, mydb. The -C option indicates that we want to create a new database (overwriting any existing database called mydb). Finally, the -v option tells af to print information about the indexing process as it goes along. (Another option, -m, may be added to increase the amount of memory used for indexing, which will generally reduce indexing time.)

The indexing process creates a group of files beginning with mydb (i.e. mydb.db, mydb.dt, and several others). These are the index files that make up the database. We can use the af tool with this database to search the files we have indexed.

Note: As an optional step after indexing, the database can be linearized. In the example above, this would be done with the command, af -L -d mydb. Linearizing improves search time but has the disadvantage of being a very slow process. Another disadvantage is that it fixes the database so that no more documents can be added.