STRING and INSPECT Statements

STRING concatenates two or more source fields or literals into a single receiving field, with each source optionally qualified by a DELIMITED BY clause that controls how much of that source is copied: DELIMITED BY SIZE copies the entire field including trailing spaces, while DELIMITED BY SPACE or a specific literal copies only up to the first occurrence of that delimiter, which is essential for concatenating fixed-length alphanumeric fields without pulling in unwanted padding. A WITH POINTER phrase lets the program control exactly where in the receiving field the next piece gets written, which is necessary when building a string across multiple STRING statements or when starting the concatenation partway through an already-populated field. If the receiving field is too small to hold everything being strung together, the excess characters are silently discarded unless an ON OVERFLOW clause is present to catch the condition.

UNSTRING performs the inverse operation of STRING, splitting a single source field into multiple receiving fields based on one or more DELIMITED BY separators, such as a comma or a specific literal, and is the standard COBOL technique for parsing delimited input like a CSV record without writing manual character-by-character scanning logic. Each receiving field in the INTO clause can optionally have a matching DELIMITER IN and COUNT IN field to capture which delimiter was actually matched and how many characters were placed into that receiving field, which is useful when the input format is not perfectly uniform. Like STRING, UNSTRING supports WITH POINTER to control the starting scan position and TALLYING to count how many receiving fields were populated, and it silently stops filling fields once either the source is exhausted or all receiving fields are used, so the ON OVERFLOW clause should be checked whenever the input might contain more delimited pieces than there are receiving fields.

INSPECT WS-FIELD TALLYING WS-COUNT FOR ALL "A" counts occurrences of a character or substring within a field without modifying the field itself, which is the standard COBOL idiom for validation tasks like counting how many non-space characters a field contains, or how many times a delimiter appears before deciding how to parse it with UNSTRING. TALLYING supports several counting modes: FOR ALL counts every occurrence, FOR LEADING counts only a contiguous run at the start of the field, and FOR CHARACTERS counts every single character position regardless of value, and multiple TALLYING phrases can appear in one INSPECT statement to gather several counts in a single pass over the field. Because INSPECT TALLYING scans the field exactly once regardless of how many counting phrases are specified, it's generally more efficient than writing an equivalent PERFORM VARYING loop that examines each character position manually.

INSPECT WS-FIELD REPLACING ALL "0" BY "O" modifies a field in place by substituting every occurrence of one character or substring with another, following the same ALL, LEADING, and FIRST qualifiers available to TALLYING, and a single INSPECT statement can combine TALLYING and REPLACING phrases to count and modify a field in one pass. INSPECT ... CONVERTING is a more restricted but often more convenient form for simple character-set translation, such as converting all lowercase letters to uppercase with INSPECT WS-FIELD CONVERTING "abcdefghijklmnopqrstuvwxyz" TO "ABCDEFGHIJKLMNOPQRSTUVWXYZ", which maps each character in the first literal to the corresponding character at the same position in the second literal across the entire field in a single statement. Unlike STRING and UNSTRING, INSPECT never changes the length of a field: REPLACING and CONVERTING only substitute characters position-for-position within the field's existing fixed size, which makes INSPECT unsuitable for any transformation that would need to insert or remove characters.