Built-in Support for PDS Processing

Currently, in order to read all members of a PDS it's necessary to first read the directory and create an array of member names, then either use an Assembler program to read each member or from PL/I dynamically allocate each fully qualified member, open the dataset, read each record from the member, then close the dataset and repeat for each member.

I'd like to see a better built-in method for processing PDS/PDSE's.

My thought is this:

DCL MYFILE FILE OUTPUT RECORD PARTITIONED; /* PARTITIONED ? New attribute available for read/write/update, the DD in the JCL/CLIST/REXX would NOT have a member name specified, just the PDS name */

ON ENDPDS(MYFILE) /* ENDPDS ? New BUILTIN function)
BEGIN;
EOPDS_MYFILE = ON;
END;

ON ENDMEMBER(MYFILE) /* ENDMEMBER ? New BUILTIN function)
BEGIN;
EOM_MYFILE = ON;
END:

OPEN FILE(MYFILE);

POSITION FILE(MYFILE) MEMBER(); /* POSITION ? New PL/I statement, MEMBER with an optional parameter for member name ? if set, point to the member specified, otherwise point to the first/next member */
DO WHILE(^EOPDS_MYFILE);
MEMBER_NAME = GETMEMBERNAME(MYFILE); /* GETMEMBERNAME ? new BUILTIN function */
EOF_MYFILE = OFF;
RECORDS_READ = 0;
READ FILE(MYFILE) INTO(RECORD_BUFFER);
DO WHILE(^EOM_MYFILE);
RECORDS_READ = RECORDS_READ + 1
READ FILE(MYFILE) INTO(RECORD_BUFFER);
END;
PUT SKIP LIST(?RECORDS READ FROM ?,MEMBER_NAME, ? = ',RECORDS_READ);
MEMBERS_PROCESSED = MEMBERS_PROCESSED + 1;
POSITION FILE(MYFILE) MEMBER();
END;

CLOSE FILE(MYFILE)

PUT SKIP LIST(?MEMBERS PROCESSED IN MYFILE = ', MEMBERS_PROCESSED)

Sample for one member named TEST:

OPEN FILE(MYFILE);

POSITION FILE(MYFILE) MEMBER('TEST');
DO WHILE(^EOPDS_MYFILE);
READ FILE(MYFILE) INTO(RECORD BUFFER);
DO WHILE(^EOM_MYFILE);
RECORDS_READ = RECORDS_READ + 1
READ FILE(MYFILE) INTO(RECORD BUFFER);
END;
PUT SKIP LIST(?RECORDS READ =',RECORDS_READ);
END;

CLOSE FILE(MYFILE)

I don't know if it's possible, but if so all of this would be done without opening and closing the dataset between members for super fast I/O.