SORTINIT

Syntax

            I16A        I16A         I16V
  SORTINIT (inputfiles, outputfiles, outputoption,
            I16V       I32V     I16V     I16A
            reclength, numrecs, numkeys, keys,
            I16A    PROC        PROC       I16A
            altseq, keycompare, errorproc, statistics,
            I16      I16        I16              I16A
            failure, errorparm, spaceallocation, charseq);

Parameters

inputfiles

16-bit signed integer array (optional)

Passes the file identification numbers of the input files to be sorted. The last element of this array should be set to zero. If the files are opened with either the NOBUF or MR (multirecord) access option (aoption), the SORT program performs the buffering and blocking/deblocking. $NULL is not a valid input file.

outputfiles

16-bit signed integer array (optional)

Passes the file identification number of the output file. The second array element must be a zero, indicating the end of the array. If the files are opened with either the NOBUF or MR (multirecord) access option (aoption), the SORT program performs the buffering and blocking/deblocking.

outputoption

16-bit signed integer by value (optional)

Passes a value that determines the format of the output records:

Value	Meaning>
0	Output record is the same as the input record (default).
1	Output record is a 32-bit integer whose value is the logical (relative) record number of the record.
2	Output record contains only the key fields, concatenated together with the major keys on the left, followed by the remaining keys.
3	Output record is the logical record number followed by the key fields.

reclength

16-bit signed integer by value (optional)

Passes the maximum length of a record in characters. If you do not specify reclength, the record length is taken from the first file specified in the inputfiles parameter. If this parameter is not specified, the inputfiles parameter must be specified.

numrecs

32-bit signed integer by value (optional)

Passes the upper bound to the number of records to be sorted. If numrecs is not specified (or if all input files are not on the disk), the value of 10,000 (32-bit integer) is used. Otherwise, the value of this parameter is derived from the file label (the end of file number of the input files).

numkeys

16-bit signed integer by value (optional)

Passes the number of keys used during the comparison of records. This parameter can be >= 1. If the numkeys parameter is specified, the keys parameter must be specified and the keycompare parameter cannot be specified. Together, numkeys and keys describe the way records are sorted.

keys

16-bit integer array (optional)

Passes information about the keys used during comparison of records. If you specify the keys parameter, you must specify the numkeys parameter and must not specify keycompare. Together, keys and numkeys describe the way records are merged.

The first element gives the position of the first character of the key within the record. The second element gives the number of characters in the key. Bits (0:8) of the third element give the ordering sequence of the records (0 for ascending, 1 for descending). Bits (8:8) of the third element indicate the type of data according to the following convention:

Value	Meaning>
0	8-bit unsigned integer
1	Twos complement (including 16-bit and 32-bit signed integer)
2	Floating-point (including real and long)
3	Packed decimal
4	Display trailing sign
5	Packed decimal with even number of digits
6	Display leading sign
7	Display leading sign separate
8	Display trailing sign separate
9	Character (using the collating sequence of charseq)
11	Short floating-point decimal
12	Floating-point decimal

NOTE: The integrity of the keys array must be maintained throughout the SORT operation. Do not change it until after the SORTEND intrinsic has been called.

altseq

16-bit signed integer array (optional)

Passes an alternate collating sequence. Bits (0:8) of the first element are specified according to the following table, where the sequence comprises the columns and the data comprises the rows:

	ASCII	EBCDIC	ALTSEQ
ASCII	chr(255)	chr(2)	chr(0)
EBCDIC	chr(1)	chr(255)	undefined

Bits (8:8) of the first element specify one less than the total number of characters in the collating sequence (in this instance, chr(255) or %377).

The remaining array elements comprise the actual collating sequence responsible for the particular SORT operation.

keycompare

procedure (reserved)

Do not specify this parameter, but maintain parameter position.

errorproc

procedure (reserved)

Do not specify this parameter, but maintain parameter position.

statistics

16-bit signed integer array (optional)

Returns information on the SORT operation, as follows:

Value	Meaning>
1, 2	Number of records sorted (32-bit integer)
3	Number of intermediate passes
4	Space available for sorting
5, 6	Number of comparisons (32-bit integer)
7, 8	Number of scratch file I/Os (32-bit integer)
9, 10	CPU time used in milliseconds (32-bit integer)
11, 12	Elapsed time in milliseconds (32-bit integer)

failure

16-bit signed integer (optional)

Returns a true value (-1) if a fatal error occurs; otherwise returns a false value (0). The value is set after each call to SORTINPUT and SORTOUTPUT. Refer to the discussion of condition codes.

errorparm

16-bit signed integer (optional)

Returns the SORTLIB error number if an error occurs. Use the SORTERRORMESS intrinsic to obtain the error message text.

spaceallocation

16-bit signed integer (reserved)

Do not specify this parameter, but maintain parameter position.

charseq

16-bit signed integer array (optional)

Passes language information in a two-element array. Set the first element to 1. Set the second element to the language ID number of the native language whose collating sequence is to be used to sort keys of type 9 (character).

Condition Codes

Related Information