Module 4: REXX Built-in Functions

REXX- Built-in Functions

REXX ships with roughly seventy built-in functions, and the string and word functions do the heavy lifting in most execs. Function names are case-insensitive, arguments are expressions, and every example below shows the real result. This page covers the functions you will use constantly: SUBSTR, WORD/WORDS, LENGTH, DATATYPE, POS, STRIP, TRANSLATE, DATE and TIME.

SUBSTR - extracting parts of a string

SUBSTR(string, n [, length [, pad]]) returns the substring starting at position n. Positions start at 1, not 0.
Omit length to take everything from n to the end. If n runs past the end, pad characters fill the result.

/* REXX */ SAY SUBSTR('mainframe', 5, 5) /* frame */ SAY SUBSTR('mainframe', 5) /* frame */ SAY SUBSTR('ABC', 5, 3, '*') /* *** (padded) */ rec = '10240MARGARET HOPPER DEV' SAY SUBSTR(rec, 1, 5) /* 10240 */ SAY SUBSTR(rec, 26) /* DEV */

WORD, WORDS and SUBWORD - working with words

REXX defines a word as text between blanks. WORDS(string) counts them, WORD(string, n) returns the nth.
SUBWORD(string, n [, length]) returns a run of words starting at word n.
Asking for a word past the end returns a null string - it never errors.

/* REXX */ line = ' one two three ' SAY WORDS(line) /* 3 */ SAY WORD(line, 2) /* two */ SAY SUBWORD(line, 2, 2) /* two three */ SAY WORD(line, 9) /* (null string) */ sentence = 'the quick brown fox' DO i = 1 TO WORDS(sentence) SAY i WORD(sentence, i) END

LENGTH, DATATYPE and POS

LENGTH(string) returns the character count, blanks included.
DATATYPE(string [, option]) tests what the string looks like: 'W' whole number, 'N' valid number, 'L' lowercase, 'U' uppercase, 'A' alphanumeric.
POS(needle, haystack [, start]) returns the position where needle first occurs, or 0 if it is absent.

/* REXX */ SAY LENGTH('z/OS') /* 4 */ SAY DATATYPE('123', 'W') /* 1 (whole number) */ SAY DATATYPE('12.5', 'W') /* 0 (not whole) */ SAY DATATYPE('12.5', 'N') /* 1 (valid number) */ SAY DATATYPE('abc', 'L') /* 1 (lowercase) */ SAY POS('world', 'hello world') /* 7 */ SAY POS('z', 'hello world') /* 0 (not found) */ IF POS(',', csvline) > 0 THEN SAY 'Line contains a comma'

STRIP, SPACE and TRANSLATE

STRIP(string [, 'B'|'L'|'T' [, char]]) removes leading/trailing/both characters - blanks by default, or any character you name.
SPACE(string [, n [, pad]]) normalizes blanks between words to n blanks (1 by default).
TRANSLATE(string) with no tables uppercases the string; with tables it maps characters one to one.

/* REXX */ SAY STRIP(' padded ') /* padded */ SAY STRIP(' padded ', 'L') /* padded (left only) */ SAY STRIP('***abc***', 'B', '*') /* abc */ SAY SPACE('a b c') /* a b c */ SAY SPACE('a b c', 2) /* a b c */ SAY TRANSLATE('MixedCase') /* MIXEDCASE */ SAY TRANSLATE('abc', 'ABC', 'abc') /* ABC */

DATE and TIME

DATE() defaults to the normal format like 8 Oct 2026. DATE('S') gives the sortable 20261008.
DATE('E') gives European 08/10/2026; DATE('U') gives USA 10/08/2026.
TIME() defaults to 24-hour 13:04:22. TIME('C') gives civil time like 1:04 PM.
Use DATE('S') for dataset qualifiers, log stamps, and anything you will sort.

/* REXX */ SAY DATE() /* 8 Oct 2026 */ SAY DATE('S') /* 20261008 */ SAY DATE('E') /* 08/10/2026 */ SAY TIME() /* 13:04:22 */ SAY TIME('C') /* 1:04 PM */ PARSE VALUE TIME() WITH hh ':' mm ':' ss stamp = DATE('S') || '_' || hh || mm || ss SAY stamp /* 20261008_130422 */

More handy functions

LEFT/RIGHT/CENTER pad or truncate to a fixed width - ideal for reports.
REVERSE, COPIES, COMPARE, VERIFY cover the remaining everyday string jobs.
ABS, MAX, MIN, TRUNC, FORMAT, RANDOM, SIGN cover everyday arithmetic.
OVERLAY, INSERT, DELSTR, CHANGESTR, COUNTSTR edit strings in place.

/* REXX */ SAY LEFT('abc', 6, '*') /* abc*** */ SAY RIGHT('abc', 6, '0') /* 000abc */ SAY CENTER('Hi', 8) /* Hi */ SAY REVERSE('abc') /* cba */ SAY COPIES('ab', 3) /* ababab */ SAY COMPARE('abc', 'abd') /* 3 (first difference) */ SAY VERIFY('12345','0123456789') /* 0 (all valid) */ SAY ABS(-5) /* 5 */ SAY MAX(3, 9, 4) /* 9 */ SAY TRUNC(3.99) /* 3 */ SAY FORMAT(3.14159, , 2) /* 3.14 */ SAY OVERLAY('XX', 'abcdef', 3) /* abXXef */ SAY CHANGESTR('l', 'hello', 'L') /* heLLo */ SAY COUNTSTR('l', 'hello') /* 2 */

Backward

Forward

Let's understand Mainframe