|
|
|
IDCAMS works correctly for physical sequential transparent data sets.
Most IDCAMS functions work properly with CA Compress because they do not process data in the cluster. The PRINT and REPRO functions work correctly whether the data is accessed through the base cluster or an alternate index path because IDCAMS uses the correct key.
Note: The FASTVSAM product substitutes itself for IDCAMS to perform REPRO. If you use FASTVSAM 2.1 or later, add IDCAMS to the EXCLUDE-PGM-SORT exclusion table (described in Exclusion Facility in the CA Compress Reference Guide) to force FASTVSAM to use record level I/O and invoke the VSAM Transparency. FASTVSAM before Version 2.1 is not supported.
The way CA Compress Custom compression handles the key makes some IDCAMS functions behave abnormally due to mismatches between the logical key location in the expanded records and the physical key location in the compressed data set. Eliminate this mismatch by defining the area from the beginning of the record to the end of the last key as a single noncompressed field, but depending on key location compression can be greatly reduced.
The BLDINDEX function can cause problems if the key is moved. To suppress key movement, always code LABEL=EXPDT=86060 on the DD statement describing the data set.
If you define an alternate index, the KEYS parameter must reflect the location of the key data after compression, since in this case the KEYS parameter defines where in the record the key appears after it has been read. But when defining a cluster, you must specify the key location for the compressed record, since KEYS in this case describes where the keys really are in the data when the VSAM access method gets them for CA Compress.
The IMPORT and EXPORT functions can give incorrect results in some situations. When the cluster is exported, the cluster characteristics are exported with the data. The data are exported in expanded mode, not compressed, so the key location in the exported characteristics does not match the actual key location. If the data set is imported into the same cluster, there is no problem. If the exported data is imported into a cluster with another name that CA Compress does not know about, a VSAM I/O error can occur because the data does not appear to be in ascending sequence.
This parameter makes the data set look to IDCAMS like a SYSIN/SYSOUT data set. Some versions of IDCAMS incorrectly modify the DCB attributes in such cases, causing 013–34 abends and other problems. To correct this problem, you need to apply the appropriate IBM maintenance.
| Copyright © 2012 CA. All rights reserved. |
|