How to index an iBase database

Indexing an iBase database is the same as indexing any ODBC database, but because this particular question appears from time to time, here's the sequence of steps to index the EXAMPLES\SCENARIO.IBS demo database supplied with the iBase.

1.  Check ODBC on your machine

First you need to have the ODBC installed and correctly running on your machine, together with drivers that support the version of database that will be indexed. For iBase you will need the MS Access driver.

ODBC can be tested by launching New Source Wizard from the Tovek Tools Index Manager. If Show Advanced Types checkbox is checked and still there isn't the ODBC Database in the list of source types, it probably means that the ODBC is somehow corrupted. One of the more common problems is that the ODBC components are of different versions. In this case - and in many others - it can be easily fixed by installing an updated ODBC/Data Access pack that can be downloaded from www.microsoft.com.

2.  Create the ODBC data source

From Control Panel start the ODBC Manager. Create a new data source that uses MS Access (*.mdb) data driver. In the dialog box for defining the actual data source do the following: enter SCENARIO_01 as a source name; as a database, select the SCENARIO.IBS from the iBase installation (default path is <PROGRAM FILES>\iBase3\EXAMPLES\SCENARIO.IBS; do not browse for *.MDB files as these are renamed to *.IBS for iBase); change System Database from NONE and select the file SCENARIO.MDA from the same folder where SCENARIO.IBS is.

Then click the Advanced button and enter the logon information. This is the preferred way to store usernames and passwords for sources to be indexed. For this iBase demo database the username and password are identical and are in the upper-case.

3.  Test your new ODBC data source

This step is not required, but it is highly recommended when creating a new type of ODBC source for the first time. This test can be performed by any application that can read data using ODBC. One example is MS Query, installed as a part of MS Office. It is used to get external data to MS Excel and can be launched in a standalone mode as well. For more information please check your MS Office documentation.

If your ODBC data source is accessible through this kind of application, indexing should then be without problems. If, however, you cannot see some tables, fields, or values, this might indicate problems with permissions, connection, etc. These problems are usually not present when accessing the database from its native application (such as MS Access is for MDB files), but they are clearly visible when you are trying to connect your data from an external program using ODBC.

4.  Create the fulltext source

When your ODBC source is ready, create the fulltext source in Tovek Tools and index the data. From TT Index Manager create a new ODBC source and in the last step fill in the following information (or use [...] buttons to select from lists): ODBC source name is SCENARIO_01, as created in the step #2; leave username/password blank because it has already been specified for the ODBC source; choose the Event data table and Unique_Id as the primary ID, and enter * for the text field which means that all the fields of every data row will be indexed.

After you create this fulltext source answer yes when asked whether it should be indexed immediately. Unmodified demo database for iBase version 3.2.2 contained 18 records in the Event table and so an empty search in Tovek Agent against this newly created source should return 18 documents in the result list as well.

Then all in one step

When you become familiar with the previous steps it's quite easy to create and index ODBC data from the TT Index Manager only. When you create a new fulltext source, from the list of ODBC sources select * Create New ODBC Source * which will start the ODBC Administrator. After you create the ODBC source everything else can be selected as in the example above. However, this faster approach is recommended only if you know the data to be indexed well and so you can spot any problem easily.