Syntax and options of the DSN1COMP control statement

Option descriptions

To run DSN1COMP, specify one or more of the following parameters on the EXEC statement to run DSN1COMP. If you specify more than one parameter, separate each parameter by a comma. You can specify parameters in any order.

32K

Specifies that the input data set, SYSUT1, has a 32-KB page size. If you specify this option and the SYSUT1 data set does not have a 32-KB page size, DSN1COMP might produce unpredictable results.

The recommended option for performance is PAGESIZE(32K).

PAGESIZE

Specifies the page size of the input data set that is defined by SYSUT1. Available page size values are 4K, 8K, 16K, or 32K. If you specify an incorrect page size, DSN1COMP might produce unpredictable results.

If you omit PAGESIZE, DSN1COMP tries to determine the page size from the input data set. Db2 issues an error message if DSN1COMP cannot determine the input page size. This might happen if the header page is not in the input data set, or if the page size field in the header page contains an invalid page size.

If information on this value is available in the input data set header page, the header page information is the default.

DSSIZE(integer G)

Specifies the data set size, in gigabytes, for the input data set. If you omit DSSIZE, Db2 obtains the data set size from the data set header page.

If you specify DSSIZE, integer must match the DSSIZE value that was specified when the table space was defined.

LARGE

Specifies that the input data set is a table space that was defined with the LARGE option. If you specify LARGE, Db2 assumes that the data set has a 4-GB boundary.

The recommended method of specifying a table space defined with LARGE is DSSIZE(4G).

If you omit the LARGE or DSSIZE(4G) option when it is needed, or if you specify LARGE for a table space that was not defined with the LARGE option, the results from DSN1COMP are unpredictable.

If information on this value is available in the input data set header page, the header page information is the default.

NUMPARTS(integer)

Specifies the number of partitions that are associated with the input data set. Valid specifications range 1 - 4096. If you omit NUMPARTS or specify it as 0, DSN1COMP assumes that your input file is not partitioned. If you specify a number greater than 64, DSN1COMP assumes that the data set is for a partitioned table space that was defined with the LARGE option, even if the LARGE keyword is not specified.

DSN1COMP cannot always validate the NUMPARTS parameter. If you specify it incorrectly, DSN1COMP might produce unpredictable results.

DSN1COMP terminates and issues message DSN1946I when it encounters an image copy that contains multiple partitions; a compression report is issued for the first partition.

This parameter is not used if the target table space is a universal table space. DSSIZE is used instead.

This parameter is deprecated.

FREEPAGE(integer)

Specifies how often to leave a page of free space when calculating the percentage of saved pages. You must specify an integer in the range 0 to 255. If you specify 0, no pages are included as free space when DSN1COMP reports the percentage of pages saved. Otherwise, one free page is included after every n pages, where n is the specified integer.

The default value is 0.

Specify the same value that you specify for the FREEPAGE option of the SQL statement CREATE TABLESPACE or ALTER TABLESPACE.

PCTFREE(integer)

Indicates what percentage of each page to leave as free space when calculating the percentage of pages saved. You must specify an integer in the range 0 to 99. When calculating the savings, DSN1COMP allows for at least n percent of free space for each page, where n is the specified integer.

The default value is 5.

Specify the same value that you specify for the PCTFREE option of the SQL statement CREATE TABLESPACE or ALTER TABLESPACE.

FULLCOPY

Specifies that a Db2 full image copy (not a DFSMSdss concurrent copy) of your data is to be used as input. Omitting this parameter when the input is a full image copy can cause error messages or unpredictable results. If this data is partitioned, also specify the NUMPARTS parameter to identify the number of partitions.

REORG

Provides an estimate of compression savings that are comparable to the savings that the REORG utility would achieve. If this keyword is not specified, the results are similar to the compression savings that the LOAD utility would achieve.

REORG does not apply if the input data set is a LOB table space.

ROWLIMIT(integer)

Specifies the maximum number of rows to evaluate in order to provide the compression estimate. This option prevents DSN1COMP from examining every row in the input data set. Valid specifications range from 1 to 99000000.

Use this option to limit the elapsed time and processor time that DSN1COMP requires. An analysis of the first 5 to 10 MB of a table space provides a fairly representative sample of the table space for estimating compression savings. Therefore, specify a ROWLIMIT value that restricts DSN1COMP to the first 5 to 10 MB of the table space. For example, if the row length of the table space is 200 bytes, specifying ROWLIMIT(50000) causes DSN1COMP to analyze approximately 10 MB of the table space.

MAXROWS(integer)

Specifies the maximum number of rows that DSN1COMP is to consider when calculating the percentage of pages saved. You must specify an integer in the range 1 to 255.

The default value is 255.

Specify the same value that you specify for the MAXROWS option of the SQL statement CREATE TABLESPACE or ALTER TABLESPACE.

EXTNDICT(dictionary-name)

Specifies the name of an external copy of the compression dictionary that DSN1COMP produces. dictionary-name must:

• Be eight bytes
• Contain only uppercase alphanumeric characters
• Begin with an alphabetic character

The external copy of the compression dictionary is primarily for use by the IBM® Data Encryption for IMS and Db2 tool.

When EXTNDICT is specified, a DSN1DICT DD statement must be included in the JCL for running DSN1COMP.

LOB

Specifies that the input data set is a LOB table space or an image copy data set of a LOB table space. When LOB is specified, DSN1COMP will estimate the space savings and compression ratio for a LOB table space using the zEnterprise® data compression (zEDC) hardware.

If the LOB table space is not compressed, DSN1COMP will calculate the compression savings that would occur if the table space is defined using the COMPRESS YES option. DSN1COMP will collect data up to the average LOB size (or 1 MB as the maximum size) and pass the collected data to the zEDC card (if zEDC hardware is installed), which will return the compressed information and statistics for the DSN1COMP report.

If the LOB table space is already compressed (the table space is defined with the COMPRESS YES option), DSN1COMP will collect statistics of the individual LOB metadata from the LOB map page without collecting and compressing data using the zEDC card again.

If you specify LOB, you cannot specify any other DSN1COMP options.

COMPTYPE

Specifies the type of compression for which DSN1COMP is to estimate the space savings. You can specify one of the following values:

HUFFMAN

Huffman compression. You can specify HUFFMAN for only universal table spaces.

FIXED

Fixed-length compression

ALL

Both Huffman compression and fixed-length compression

If you do not specify the COMPTYPE option, the compression type for which DSN1COMP provides space estimates depends on the hardware. If DSN1COMP is run on IBM z14® or later hardware, the utility provides estimates for both Huffman compression and fixed-length compression. Otherwise, DSN1COMP provides estimates for fixed-length compression only.

Related information:

• Using Huffman compression to compress your data
• Using fixed-length compression to compress your data

LEAFLIM(integer)

Specifies how many index leaf pages should be evaluated to determine the compression estimate. This option prevents DSN1COMP from processing all index leaf pages in the input data set. Valid specifications range from 1 to 99000000.

If the LEAFLIM parameter is not specified, the entire index will be scanned and all leaf pages will be examined.

In a compressed index, only leaf pages are compressed. All other page types remain uncompressed.