VSAM Record Selection Efficiencies

In this section: Reporting From Files With Alternate Indexes

The most efficient way to retrieve selected records from a VSAM KSDS data source is by applying an IF screening test against the primary key. This results in a direct reading of the data using the data source's index. Only those records that you request are retrieved from the file. The alternative method of retrieval, the sequential read, forces the data adapter to retrieve all the records into storage.

Selection criteria that are based on the entire primary key, or on a subset of the primary key, cause direct reads using the index. A partial key is any contiguous part of the primary key beginning with the first byte.

IF selection tests performed against virtual fields can take advantage of these efficiencies as well, if the full or partial key is embedded in the virtual field.

The EQ and IS relations realize the greatest performance improvement over sequential reads. When testing on a partial key, equality logic is used to retrieve only the first segment instance of the screening value. To retrieve subsequent instances, NEXT logic is used.

Screening relations GE, FROM, FROM-TO, GT, EXCEEDS, IS-MORE-THAN, and NOT-FROM-TO all obtain some benefit from direct reads. The following example uses the index to find the record containing primary key value 66:

IF keyfield GE 66

It then continues to retrieve records by sequential processing, because VSAM stores records in ascending key sequence. The direct read is not attempted when the IF screening conditions NE, IS-NOT, CONTAINS, OMITS, LT, IS-LESS-THAN, LE, and NOT-FROM are used in the report request.