Selecting the scan method
IMS Index Builder provides a sequential scan technology for building primary, secondary, and ILDS indexes, and this sequential scan technology is used as the default scan method. However, you can change the scan method to apply the scan method that was used in IMS Index Builder 2.3.
About this task
When you build a primary index and secondary indexes together in IMS Index Builder 2.3, the primary database was scanned twice, once by using the sequential scan method and once by using the hierarchical HPU scan method. However, in this version of IMS Index Builder, the primary database is scanned only once by using the sequential scan method, which reduces the elapsed time.
When the sequential scan method is used to build one or more secondary indexes, several sort tasks are started depending on the segment level of the index sources. The segment occurrences of the primary database are read in physical order, then the index pointers and the index keys are collected and sorted in hierarchical order. For example, when segment level of an index source is the second level of dependent segments, two sort tasks (IIUSRTV1 and IIUSRTV2) are started, and when the index target or the index source is a variable-length segment, an additional sort task (IIUSRTD0) is started.
These sort tasks require additional sort workspace. The sort file size of IIUSRTVx is estimated based on the sum of index records, and the estimated value is given to the SORT FILSZ option. When the VSAM data set for the index is not reused, the number of index records is estimated based on the allocation characteristics of the VSAM data set. You can override the estimated value by using the SORTFSZ control statement.
When additional sort workspace is not available, you can use the scan method that was used in IMS Index Builder 2.3.