Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011
Posts: 7309
Location: Inside the Matrix
|
|
|
|
The new version DFSORT z/OS V2R1 has been official released as of September 30 2013. The following are list of new features of z/OS V2R1 - Base (September 30, 2013)
64-bit support
Eligible user programs and exits can now be written to:
- Call DFSORT from a program in 64-bit addressing mode (AMODE 64) using a new 64-bit parameter list and the entry name ICEMAN64 or SORT64
- Use DFSORT E15, E35 and E32 exits running in 64-bit addressing mode (AMODE 64)
- Pass 64-bit addressed records to DFSORT using new 64-bit parameter lists for E15, E32 and E35 exits.
Performance and resource usage improvements
DFSORT has been enhanced to improve use of central storage in relation to system activity. The existing EXPMAX, EXPOLD and EXPRES installation options will now be evaluated in conjunction with available resources at run-time which can give you better control over how DFSORT uses available central storage resources. The likelihood of over committed central storage resources and excessive paging has been reduced, which can provide improved reliability and performance for all workloads including DFSORT.
The shipped default for EXPOLD is changed from MAX to 50%. The shipped to system default for EXPRES is changed from 0 to 10%.
DFSORT has been enhanced to allocate storage in smaller increments and then check resource availability before each additional increment is reserved. This can allow concurrent sort applications to better share available storage and allows sorts to stop allocating additional storage if resources are no longer available. Additionally, DFSORT's dynamic allocation of work data sets will be adjusted to reduce the likelihood of failure when a sort application is unable to obtain all of the expected central storage and must then use more disk work space than expected.
A new TUNE installation default allows you to specify whether DFSORT should allocate storage in increments with additional disk work space to minimize the risk of failure, or to allocate all storage at initialization so disk work space allocation can be reduced.
Blockset merge applications can now use storage above 16 MB virtual with more functions (such as E61, INREC, OUTREC, INCLUDE, OMIT and SUM), providing improved performance and virtual storage constraint relief.
XTIOT, uncaptured UCBs and DSAB above 16 megabytes virtual
Work data sets dynamically allocated by DFSORT will use options for XTIOT, uncaptured UCBs, and DSAB above 16 megabyte virtual to the extent that z/OS supports them.
Extended address volumes
DFSORT support for EAS-eligible data set types on Extended Address Volumes has been enhanced to increase the maximum size of a DFSORT work data set. With full track blocking, the maximum number of tracks that can be used for a single work data set has been increased from 1048576 to 16777215. In situations where DFSORT must use a reduced block size for the work data sets, less than 16777215 tracks can be used.
Alphanumeric comparison tests
New UC, LC, MC, UN, LN and MN keywords (similar to the previously available NUM keyword) now allow you to test a field for various combinations of alphanumeric characters or non-alphanumeric characters using binary (BI) format.
The new keywords allow you to select any of the following alphanumeric character sets:
- UC: Uppercase characters (A-Z)
- LC: Lowercase characters (a-z)
- MC: Mixed case characters (A-Z, a-z)
- UN: Uppercase and numeric characters (A-Z, 0-9)
- LN: Lowercase and numeric characters (a-z, 0-9)
- MN: Mixed case and numeric characters (A-Z, a-z, 0-9)
You can use these new keywords in the following comparison operands: COND, INCLUDE, OMIT, BEGIN, END, WHEN and TRLID.
More PARSE fields
You can now use up to 1000 parsed fields (%0-%999) with the PARSE function; the previous limit was 100 parsed fields (%0-%99).
Repeating PARSE fields
REPEAT=v is a new PARSE option that can be used to repeat a particular parse field definition multiple times.
REPEAT=v can be used with % to specify v identically defined consecutive parsed fields to be ignored.
REPEAT=v can be used with %n, %nn or %nnn to specify v identically defined consecutive parsed fields for which data is to be extracted. The parsed fields will start with the %n, %nn or %nnn field you select and be incremented by one for each repeated parsed field.
Alphanumeric tests for PARSE fields
STARTAFT=an, STARTAT=an, ENDBEFR=an and ENDAT=an can now be used with the PARSE function to start or end when a character from any of various alphanumeric character sets is found. New keywords for an allow you to select any of the following alphanumeric character sets:
- UC: Uppercase characters (A-Z)
- LC: Lowercase characters (a-z)
- MC: Mixed case characters (A-Z, a-z)
- UN: Uppercase and numeric characters (A-Z, 0-9)
- LN: Lowercase and numeric characters (a-z, 0-9)
- MN: Mixed case and numeric characters (A-Z, a-z, 0-9)
- NUM: Numeric characters (0-9)
Insert a string at the end of variable-length records
VLTRAIL=string is a new OUTFIL option that allows you to insert a character string (C'string') or hexadecimal string (X'yy...yy') at the end of each variable-length OUTFIL output record.
Symbol enhancements
DFSORT Symbols can now be used for many more operands. The following operands of the form KEYWORD=n can now be specified as KEYWORD=symbol where symbol represents an equivalent number (for example, if you have New_Length,25 in SYMNAMES, you can use LENGTH=New_Length wherever you can use LENGTH=25): ABSPOS, ACCEPT, ADDPOS, AVGRLEN, DO, ENDPOS, ENDREC, FIXLEN, ID, IFOUTLEN, INCR, LENGTH, LINES, MAXLEN, RECORDS, REPEAT, SAMPLE, SEQ, SKIPREC, SPLIT1R, SPLITBY, START, STARTPOS, STARTREC, STOPAFT and SUBPOS.
A symbol for a number can now be used for the length of the output field with the CHANGE operand, and for the length of the sequence number with the SEQNUM operand.
A symbol can be used for any of the new %000-%999 parsed fields.
A symbol can be used for string with the new VLTRAIL=string operand.
Improved diagnostics
DFSORT now provides specific reason codes and associated documentation to aid in diagnosing and correcting errors associated with messages ICE083A and ICE141A.
The text for message ICE118I has been changed to recommend the use of FILSZ=En.
The ICE236I options-in-effect message now includes values for TUNE, EXPMAX, EXPOLD and EXPRES.
The text for message ICE285A has been changed for clarification.
Check this link for publications and complete details
Use [URL] BBCode for External Links |
|
