Recommendations for defining Alternative Indexes

The use of Alternative indexes requires additional resources during User file updating, so Alternative indexes should be defined carefully to fit retrieval requests and achieve the desired aim.

Consider the following factors when choosing an index to be used:

• Number of index records to be retrieved to build the requested report list.
• Number of index records that will be rejected.
• Number of rejected user records.

The fields that can be used in Alternative indexes are listed in the description of the CTDUFANX utility, and can be divided into three groups:

• Fields that are specified as fixed values in selection criteria, such as Report Name, Job Name, Category, and so on.
• Date fields, which are specified as range in selection criteria.
• Recipient name, which can be multiplied, if SYNONYM or descendants are used in the recipient tree. This field is mandatory in an index key, because the recipient should always be checked against the recipient tree for security purposes.

To decrease the number of retrieved index records, the Alternative Index key should be composed of fields that are frequently specified in the selection criteria.

To make the index more effective, the first fields of the key should be fields of the group 1, if they are fully specified in the filter. The following field should be either a field from group 1, which is specified as prefix or a field from groups 2 or 3, depending on the implementation.

The fields that are specified as masks should be positioned at the end of Alternative Index key.

The order of the fields in the Alternative Index key should be defined so that the search process will have to go through the smallest number of records.