Adabas Sequential Files

This document describes Adabas sequential files.


Sequential File Table

This section summarizes the sequential files used by the Adabas utilities. Explanations of the table heading and contents are in the text following the table.

Utility File Name Out In BLKSIZE
by device
Concatenation
ADACDC DDSIIN   x   Yes
ADACHK DDFEHL x   Yes  
ADACMP DDAUSBA
DDEBAND
DDFEHL
x
x
x   Yes
ADACNV DDFILEA x      
ADALOD DDEBAND
DDFILEA
DDISN
DD/OLD
x
x
x
x
x
Yes Yes
Yes
ADAMER DDEBAND   x    
ADAORD DDFILEA x x Yes  
ADAPLP DD/PLOG   x   Yes
ADARAI DDOUT x      
ADAREP DDSAVE
DDPLOG
  x
x
  Yes
Yes
ADARES DDBACK
DDSIAUS1
DDSIAUS2
DDSIIN
x
x
x
x
  Yes
Yes
ADASAV DDDEL1
DDDEL2
DDDEL3
DDDEL4
  x
x
x
x
  Yes
Yes
Yes
Yes
DDDEL5
DDDEL6
DDDEL7
DDDEL8
  x
x
x
x
  Yes
Yes
Yes
Yes
DDDUAL1
DDDUAL2
DDDUAL3
DDDUAL4
x
x
x
x
     
DDDUAL5
DDDUAL6
DDDUAL7
DDDUAL8
x
x
x
x
     
DDFULL
DDPLOG
  x
x
  Yes
Yes
DDREST1
DDREST2
DDREST3
DDREST4
  x
x
x
x
  Yes
DDREST5
DDREST6
DDREST7
DDREST8
  x
x
x
x
   
DDSAVE1
DDSAVE2
DDSAVE3
DDSAVE4
x
x
x
x
     
DDSAVE5
DDSAVE6
DDSAVE7
DDSAVE8
x
x
x
x
     
ADASEL DDEXPA1
DDEXPA2
DDEXPA3
DDEXPA4
x
x
x
x
     
DDEXPA5
DDEXPA6
DDEXPA7
DDEXPA8
x
x
x
x
     
DDEXPA9
DDEXPA10
DDEXPA11
DDEXPA12
x
x
x
x
     
DDEXPA13
DDEXPA14
DDEXPA15
DDEXPA16
x
x
x
x
     
DDEXPA17
DDEXPA18
DDEXPA19
DDEXPA20
x
x
x
x
     
DDSIIN   x   Yes
DDEBAND   x   No
DDSAVE   x   No
ADAULD DDOUT1
DDOUT2
DDISN
DDSAVE
DDPLOG
DDFULL
DDDEL1-8
x
x
x
x
x
x
x
Yes
Yes
Yes
Yes
Yes
Yes
Yes

Files that are both output and input are first written and then read by the indicated program.

z/OS Dependencies

To allow utilities to access data set information after closing, the DD statement for sequential data sets used in utilities should not contain FREE=CLOSE.

The BLKSIZE of a sequential file is determined as follows:

  • If the column, "BLKSIZE by device" specifies "Yes" for a file, the default BLKSIZE depends on the device type as follows:

    Tape: up to 256K
    3380 disk: 23476
    3390 disk: 27998
  • If the column "BLKSIZE by device" does not specify "Yes" for a file, the file's BLKSIZE is obtained from the DD statement or data set label, if present. It must be present for any input file.

  • If the column "BLKSIZE by device" does not specify "Yes" for a file and the BLKSIZE cannot be obtained from the DD statement or data set label, the value of the ADARUN QBLKSIZE parameter is used, if specified.

Except for ADACMP EBAND, the RECFM and LRECL of all sequential files are VB and BLKSIZE-4, respectively. For ADACMP EBAND, RECFM and LRECL must be available from the DD statement and/or data set label.

If the DCB BUFNO parameter is provided on the DD statement, it will be used.

If the DCB BUFNO parameter is not provided on the DD statement, 40 buffers are allocated for disk and tape datasets with a block size of less than 64K, and 10 buffers are allocated for tape datasets with a block size of 64K or larger.

Enhancements to TEMP and SORT Processing in ADAINV, ADALOD and ADAULD

Various enhancements to the processing of TEMP and SORT datasets are now available in utilities ADAINV, ADALOD and ADAULD. Similar enhancements are already available for use by the ADACHK utility.

  • Allow SORT and TEMP to be dynamically extended when full.

  • Support writing to SORT and TEMP even though the datasets have not been pre-formatted (e.g. via ADAFRM).

  • Support temporary dataset usage for TEMP and SORT, e.g.:

    //DDTEMPR1 DD  DSN=&&TEMP,DISP=(NEW,DELETE),
    //             SPACE=(CYL,(100,100)),VOL=(,,,10)
    //DDSORTR1 DD  DSN=&&SORT,DISP=NEW,
    //             SPACE=(CYL,(100,100)),RECFM=F,BLKSIZE=8904

The TEMPSIZE and SORTSIZE parameters, which were previously mandatory, are now optional parameters.

If TEMPSIZE/SORTSIZE is specified, the utility will process TEMP/SORT in the same way as in previous versions, i.e. pre-formatting of TEMP/SORT is required and dynamically extending the dataset is not available.

By omitting the TEMPSIZE/SORTSIZE parameter, users can take advantage of the enhancements listed above.

When TEMPSIZE/SORTSIZE is omitted, the user can either let the utility choose the TEMP/SORT device type (by omitting the TEMPDEV/SORTDEV parameter), or else set the TEMP/SORT device type by specifying TEMPDEV/SORTDEV.

When both TEMPSIZE/SORTSIZE and TEMPDEV/SORTDEV are omitted and TEMP/SORT is unformatted, users can specify the block size to be used for TEMP/SORT via JCL. The sample JCL for DDSORTR1 above provides an example of this. In this case, the block size in the JCL will be used if it matches an Adabas device type. A JCL block size which does not match an Adabas device type is treated as an error. For the list of Adabas device types and corresponding block sizes, see Supported Adabas Device Types.

For processing TEMP/SORT datasets which are not pre-formatted, Software AG recommends omitting both TEMPSIZE/SORTSIZE and TEMPDEV/SORTDEV, and not specifying a block size via JCL. The utility will then select an optimum block size.

When both TEMPSIZE/SORTSIZE and TEMPDEV/SORTDEV are omitted and TEMP/SORT is pre-formatted, the utility retains the block size of the existing dataset.

Note that it is still possible for TEMP/SORT to fill up. Users should define TEMP and SORT datasets with secondary extents large enough to hold all required data.

In general, TEMP must be large enough to accommodate all descriptor values to be loaded into the Normal Index, and SORT must be large enough for twice the largest size of all values for any one descriptor to be sorted.

See the , and Utilities documentation for further guidance on the size requirements for TEMP/SORT datasets.