I have been coding in COBOL for many years, but until recently I did not have to use the functionality provided by INDEX in PL/I, POS in REXX and POSSTR in SQL. COBOL is equipped with a very limited set of functions and not one of them looks like the mentioned ones. I had to ask more experienced COBOL programmers for help and I was told that INSPECT might do the trick.

The descriptions in the COBOL manual are not famous for readability, so we had to find an example before I could create some usable code. I hope this tip may save you some time compared to the frustrations I had to go through. Just to find out that INSPECT is the COBOL statement that will help is not exactly intuitive. And here is how to do it:

01 needle PIC X(5).
01 haystack PIC X(50).
01 needlePos PIC S9(9) BINARY.

MOVE 'COBOL' TO needle
MOVE 'ITISNOTEASYTOCODEINDEXORPOSORPOSSTRINCOBOLIDOTHINK' TO haystack

MOVE 1 TO needlePos
INSPECT haystack TALLYING needlePos FOR CHARACTERS BEFORE needle

IF needlePos > LENGTH OF haystack
  DISPLAY 'NEEDLE ' needle ' NOT FOUND IN HAYSTACK ' haystack
ELSE
  DISPLAY 'NEEDLE ' needle ' FOUND IN POSITION ' needlePos
END-IF

INSPECT counts the number of bytes before the first occurence of NEEDLE and adds the count to NEEDLEPOS. That is why I start by initializing NEEDLEPOS with 1 before executing the INSPECT statement. If NEEDLE is not found in HAYSTACK then NEEDLEPOS ends up by being the number of bytes in HAYSTACK plus one, very unlike INDEX/POS/POSSTR where NEEDLEPOS would be set to zero. By the way INSPECT is capable of many other things than just scanning strings.