Database Detail
When you first create a profile, Sawmill will ask you what kind of information you want to track in your profile. The values you choose determine what your initial profile settings are, including the database cross-references and the available views.
The options available depend in your log format, but may include some of the following:
Track all fields day-by-day. Turning this option cross-references the date/time field to all other fields. This improves performance when the Calendar or Date Range controls are used. Day-by-day information will still be available in the database if this option is not checked, but full table scans will be required to query it, which may make the queries much slower.
Track hosts individually. Turning this option on structures the "host," "browsing host," or "source IP" field so that all IPs are tracked fully. If you leave this off, Sawmill will track only the top level domains (e.g. yahoo.com) and subnets (e.g. 127.128) of the IP's, and you will not be able to get information about the activities of a particular IP. If you turn this on, every IP address in the log data will be tracked separately, so information will be available about individual IP addresses and hostnames. Turning this on can significantly increase the size and memory usage of this database.
In addition, there will be a checkbox for each available numerical field in the log data. Checking one of these boxes will add another field to the database, providing information about that numerical field, and will add that numerical field to every report. This will slightly increase the size of the database for most fields, but tracking a "unique" field like visitors may be much more expensive. Turning on unique host (visitor) tracking will result in the visitor id information being tracked for all database items, which will significantly slow log processing and increase database size, but it is necessary if you need visitor information. For web and web proxy logs, you can greatly increase processing speed (as much as four times) by checking only the "page views" box (and not track hits or bandwidth).