User`s guide

23.8.2 Considerations for using Saved Data Indexes
There are some things that you should consider before deciding to use Saved Data Indexes:
• Indexes work best in scenarios where a report contains a large set of data, but only small subsets
of the data are viewed at one time.
• Large, complex reports can overload the Crystal Reports report engine.
For example, creating a large report with a selection formula that returns 90% of the report's data
will probably take up more memory and disk space than a report file without indexes. In such a case,
report processing will be slower.
• Indexes require disk space; this disk space can add up to more than the space required for the data
itself. You trade the speed of applying selection formulas for the size of your report files.
• The Crystal Page and RAS servers are limited in how much caching occurs for different users when
selection formulas are applied. Therefore, when a user views a report file with a selection formula
and the Page Server takes up a certain amount of RAM, adding more users could cause the Page
Server to use even more RAM.
23.8.3 Indexing the right field(s)
These guidelines describe the best ways to index saved data and what to avoid when indexing:
• Index fields that users frequently add to their record selection formulas.
• Index fields that are referred to by the report's record selection formula.
• Don't index all of the fields in the report.
Doing so can result in increased processing times. It is best to index only on the fields that meet the
criteria specified above. If all of the fields meet these criteria, then you should prioritize the fields
and index only some of them.
• Don't index fields that contain unique values only.
For instance, don't index a field such as "Last Year's Sales," whose values are likely to be distinct
from one another. If you do so, a separate index is created for each and every value in the field.
23.8.3.1 To index saved data
2012-03-14552
Understanding Databases