The TUNE INDEX utility output indicates a count for the number of records read as well as the number of SR8s processed. These counts displayed at the end of the execution can be higher than the number of records contained in the index set.
Component: IDMS - All supported releases
This is because of the method used by the index to achieve its results.
TUNE INDEX follows the sequence of first performing orphan adoption, then re-balancing, and finally re-sequencing. Orphan adoption is relatively straight-forward, but the other two tend to cause SR8s to be moved around within the area.
Through general orphan adoption, TUNE INDEX locates an index and processes it through the SR8s moving bottom to top, left to right. As level-0 is processed, the data record is accessed to update the UP pointer. On intermediate levels, lower level SR8 are accessed to adjust their UP pointers. This orphan adoption is done across an entire level of the index at one time.
During re-balance and re-sequence processing, individual SR8 records are accessed and following any splits the orphan adoption routine is called to process a single SR8. In these cases, data records may again be referenced to update UP pointers. Most of those numbers in the reports are fairly meaningless to a user to determine any correlation to the size of the index.
Each time the utility needs to read an SR8, the counter is updated. Re-sequencing is especially work intensive as SR8s may be moved off of a page so that other SR8s can be moved onto that page based on the sequence of the index. This can result in a ripple effect across the index's structure causing a great increase in the number of SR8 records read.