Version 8.1.4	SEARCH   INDEX   CONTENTS   |   PDF PAGE   PDF BOOKS   |   HOME   UP   PREV   NEXT

This document describes the ADACMP COMPRESS function output.

• Compressed Data Records

• Rejected Data Records

• ADACMP Report

Compressed Data Records

The data records that ADACMP has processed, edited, and compressed are written out together with the file definition information to a sequential data set with the variable blocked record format. This data set may be used as input to the ADALOD utility. The output of several ADACMP executions may also be used as input to ADALOD.

If the output data set contains no records (no records provided on the input data set or all records rejected), the output may still be used as input to the ADALOD utility. In this case, you must ensure that the amount of Associator space allocated to the file is sufficient since an accurate estimate cannot be made by the ADALOD utility without a representative sample of input record values (see the ADALOD utility for additional information).

For information on how to identify MU and PE occurrences greater than 191 in the compressed record, read Identifying MU and PE Occurrences Greater Than 191 in Compressed Records.

Rejected Data Records

Any records rejected during ADACMP compression are written to the DD/FEHL error data set. The records are output in variable blocked format and may be segmented into multiple physical records. Each logical rejected record will be preceded by an initial ADAF rejected record header. If the logical record and the ADAF header do not fit in the DD/FEHL physical record length, ADAN rejected record headers will precede the remaining physical rejected record segments that comprise the logical rejected record.

The functions of these two different headers are as follows:

• ADAF headers indicate the error condition and pertinent information.

• ADAN headers are smaller and are used for rejected record continuation and ADAH/ADAC header error reporting.

DSECTs for the ADAF and ADAN headers can be found in members ADAF and ADAN of the distributed Adabas SRCE data set.

Traditionally, the DD/FEHL error data set produced for ADACMP errors has truncated rejected records that exceeded the FEHL physical record length. In Version 8, the rejected records are segmented instead of truncated. Because of this change, the DD/FEHL LRECL setting must be at least 500 bytes.

If HEADER=YES is specified, an error may occur while segmented uncompressed records are being assembled into a logical record. If the ADAH header is in error, the ADAH record is written and subsequent ADAC records are not written until the next ADAH record is processed. If an ADAC header is in error, the preceding ADAH header will be written without its payload data. The ADAC record in error will be written in its entirety. Subsequent ADAC records are not written until the next ADAH record is processed.

The following response codes may occur:

Only the first incorrect field within a record is detected and referenced. If there are other errors, they are not detected until subsequent runs are made.

Example of Rejected Data Records

The following table depicts the FEHL output for four rejected records during ADACMP compression. Rejected records 1 and 3 have only one FEHL record (ADAF); rejected record 2 is segmented into two FEHL records (ADAF and ADAN); rejected record 4 is segmented into three FEHL records (one ADAF record and two ADAN records):

Note:
DSECTs for the ADAF and ADAN headers can be found in members ADAF and ADAN of the distributed Adabas SRCE data set.

ADACMP Report

ADACMP calculates the approximate amount of space (in both blocks and cylinders) required for Data Storage for the compressed records. This information is printed as a matrix which contains the required space for the different device types requested by the DEVICE parameter for various Data Storage padding factors between 5 and 30 percent.

The following is an example of ADACMP report output:

PARAMETERS:
   
ADACMP COMPRESS NUMREC=1000
ADACMP FNDEF='01,AA,8,B,DE'
ADACMP FNDEF='01,BA,6,A,NU'
ADACMP FNDEF='01,BB,8,P,NU'
ADACMP FNDEF='01,AD,1,A,FI'
ADACMP SUBDE='CA=BA(1,3)'

COMPRESS PROCESSING STATISTICS:

NUMBER OF RECORDS READ                1,000
NUMBER OF INCORRECT RECORDS               0
NUMBER OF COMPRESSED RECORDS          1,000

RAW DATA                             24,000 BYTES
COMPRESSED DATA                      16,656 BYTES
COMPRESSION RATE                       31.9 %
LARGEST COMPRESSED RECORD                20 BYTES


DATASTORAGE SPACE REQUIREMENTS:

I  DEVICE  I  PADDING I  BLOCKSIZE I          NUMBER OF     I
I          I  FACTOR  I    BYTES   I      BLOCKS      CYLS  I
I----------I----------I------------I------------------------I
I   3380   I          I    4,820   I                        I
I          I     5%   I    4,578   I           4         1  I
I          I    10%   I    4,337   I           4         1  I
I          I    15%   I    4,096   I           5         1  I
I          I    20%   I    3,856   I           5         1  I
I          I    25%   I    3,615   I           5         1  I
I          I    30%   I    3,373   I           5         1  I
I          I          I            I                        I
I----------I----------I------------I------------------------I

TEMP SPACE ESTIMATION:

I  DEVICE  I  BLOCKSIZE I          NUMBER OF     I
I          I    BYTES   I      BLOCKS      CYLS  I
I----------I------------I------------------------I
I   3380   I    7,476   I           5          1 I
I----------I------------I------------------------I

THE LARGEST DESCRIPTOR IS AA, IT WILL OCCUPY 1 TEMP BLOCKS


SORT SPACE ESTIMATION:

I DEVICE I BLOCKSIZE I   LWP              I      NR OF      I
I        I  (BYTES)  I  (BYTES)           I  BLOCKS    CYLS I
-------------------------------------------------------------
I  3380  I     7476  I   139264 (MINIMUM) I       2       1 I
I        I           I  1048576 (DEFAULT) I       2       1 I
I        I           I   139264 (OPTIMUM) I       2       1 I
I--------I-----------I--------------------I-----------------I

The compression rate is computed based on the real amount of data used as input to the compression routine. Fields skipped by a format element "nX" (used to fill a field with blanks) are not counted.

If SPAN was specified in an ADACMP COMPRESS run, statistics about the spanned Data Storage records are also printed:

Spanned Record Statistics:                    
--------------------------                    
                                              
Number of Non-Spanned records               2 
Number of Spanned records                   8 
Min Number of Segments              2         
Max Number of Segments              5         
Avg Number of Segments              4         
                                              
Max MU Count                        0         
Max PE Count                      300         

If large object (LB) fields are compressed, statistics about the LB fields (listed as "LOBs" in the report) are printed:

Large Objects Statistics:                           
--------------------------                          
                                                    
Number of bytes for largest LOB           10,124,996
Total number of LOBs                                         5
Total number of outsourced LOBs                        5
Number of LOB file records                         10,085

www.softwareag.com

Copyright © 1971-2014 Software AG, Darmstadt, Germany.

SEARCH   INDEX   CONTENTS   |   PDF PAGE   PDF BOOKS   |   HOME   UP   PREV   NEXT