This is sadly not possible. The only way to do something like this would be to write a new record with the new key, and delete the old one. You would probably want to do this within a transaction, so that either both operations succeed, or they both fail.

Can we not do something like below?  Like two different data files. Then is there a command to swap the data file names?

IDENTIFICATION DIVISION.
PROGRAM-ID. CHANGE-PRIMARY-KEY.

ENVIRONMENT DIVISION.
INPUT-OUTPUT SECTION.
FILE-CONTROL.
    SELECT OLD-FILE ASSIGN TO DISK "OLD_DATA"
        ORGANIZATION IS SEQUENTIAL
        ACCESS MODE IS SEQUENTIAL.
    SELECT NEW-FILE ASSIGN TO DISK "NEW_DATA"
        ORGANIZATION IS SEQUENTIAL
        ACCESS MODE IS SEQUENTIAL. 

DATA DIVISION.
FILE SECTION.
    FD OLD-FILE
        RECORD CONTAINS 100 CHARACTERS
        DATA RECORD IS RECORD-DATA.
    FD NEW-FILE
        RECORD CONTAINS 100 CHARACTERS
        DATA RECORD IS RECORD-DATA.

DATA DIVISION.
WORKING-STORAGE SECTION.
    01 RECORD-DATA.
        05 CUSTOMER-ID PIC 9(5).
        05 CUSTOMER-NAME PIC X(30).
        05 OTHER-DATA PIC X(65). 

PROCEDURE DIVISION.
    OPEN INPUT OLD-FILE OUTPUT NEW-FILE. 
    
    READ OLD-FILE INTO RECORD-DATA AT END
        GO TO EXIT-PROCESS. 

    PERFORM UPDATE-RECORD UNTIL 
        READ OLD-FILE INTO RECORD-DATA AT END.

    CLOSE OLD-FILE NEW-FILE. 

    EXIT-PROCESS.
    EXIT.

    UPDATE-RECORD SECTION. 
        /* Logic to modify the CUSTOMER-ID (primary key) */
        COMPUTE CUSTOMER-ID = CUSTOMER-ID + 1. 
        WRITE RECORD-DATA TO NEW-FILE.