Substring a text using MOVEL function in RPG

Question

Question:

Is it save to get substring n characters from a text in RPG using MOVEL function which take a text with length x and store it to a variable with capacity n?

Or the only save way to get the first n character is using SUBST?

The background of the question is one of my colleague getting the first 3 characters from a database with 30 char in length is using MOVEL to a variable with length only 3 char (like truncating the rest of it). The strange way, sometimes the receive variable is showing minus character ('-'), sometimes doesn't. So I assume using MOVEL is not a safe way. I am thinking like string in C which always terminated by '\0', you need to use strcpy function to get the copy save, not assigning using = operator.

Anybody who knows RPG familiar with this issue?

Answer 1

MOVEL should work. RPG allows several character data types. Generally speaking, someone using MOVEL will not be dealing with null terminated strings because MOVEL is an old technique and null terminated strings are a newer data type. You can read up on the MOVEx operations and the string operations in the RPG manual. To get a better answer, please post your code, including the definitions of the variables involved.

EDIT: Example of how MOVEL handles signs.

   dcl-s long char(20) inz('CORPORATION');
   dcl-s short char(3) inz('COR');
   dcl-s numb packed(3: 0);

   // 369
 c                   movel     long          numb
   dsply numb;

   // -369
 c                   movel     short         numb
   dsply numb;


   *inlr = *on;

Answer 2

With signed numeric fields in RPG the sign is held in the zone of the last byte of the field. So 123 is X'F1F2F3' but -123 is X'F1F2D3'. If you look at those fields as character strings they will have 123 and 12L in them.

In your program you are transferring something like "123 AAAAAL" to a 3 digit numeric field so you get X'F1F2F3' but because the final character is X'D3' that changes the result to have a zone of D i.e. X'F1F2D3' You anomaly is dependent on what the 30th character contains. If it is } or any capital letter J to R then you get a negative result. [It doesn't matter whether the first 3 characters are numbers or letters because it is only the second half of the byte, the digit, that matters in your example.]

The IBM manuals say: If factor 2 is character and the result field is numeric, a minus zone is moved into the rightmost position of the result field if the zone from the rightmost position of factor 2 is a hexadecimal D (minus zone). However, if the zone from the rightmost position of factor 2 is not a hexadecimal D, a positive zone is moved into the rightmost position of the result field. Other result field positions contain only numeric characters.

Don