However, the first way provides easier program readability when working with multiple record types. For example, statements (1) and (2) in the following example are logically equivalent:

FILE SECTION. FD STOCK-FILE. 01 STOCK-RECORD PIC X(80). WORKING-STORAGE SECTION. 01 STOCK-WORK PIC X(80). ----------------(1)---------------- --------------(2)--------------- WRITE STOCK-RECORD FROM STOCK-WORK. MOVE STOCK-WORK TO STOCK-RECORD. WRITE STOCK-RECORD.

When you omit the FROM phrase, you process the records directly in the record area or buffer (for example, STOCK-RECORD).

The following example writes the record PRINT-LINE to the device assigned to that record's file, then skips three lines. At the end of the page (as specified by the LINAGE clause), it causes program control to transfer to HEADER-ROUTINE.

WRITE PRINT-LINE BEFORE ADVANCING 3 LINES AT END-OF-PAGE PERFORM HEADER-ROUTINE.

For a WRITE FROM statement, if the destination area is shorter than the file's record length, the destination area is padded on the right with spaces; if longer, the destination area is truncated on the right. This follows the rules for a group move.

6.3.3 File Handling for Relative Files

Creating a relative file involves the following tasks:

1. Specifying ORGANIZATION IS RELATIVE in the Environment Division SELECT clause
2. Specifying ACCESS MODE IS SEQUENTIAL (or RANDOM) in the Environment Division SELECT clause
   Each of these two access modes requires a different processing technique. (Refer to the Creating a Relative File in Sequential Access Mode and Creating a Relative File in Random Access Mode sections in this chapter for information about those techniques.)
3. Opening the file for OUTPUT or I-O
4. Initializing the relative key data name for each new record
5. Executing a WRITE statement for each new relative record
6. Closing the file

Creating a Relative File in Sequential Access Mode

When your program creates a relative file in sequential access mode, the I/O system does not use the relative key. Instead, it writes the first record in the file at relative record number 1, the second record at relative record number 2, and so on, until the program closes the file. If you use the RELATIVE KEY IS clause, the compiler moves the relative record number of the record being written to the relative key data item. Example 6-24 writes 10 records with relative record numbers 1 to 10.

Example 6-24 Creating a Relative File in Sequential Access Mode

IDENTIFICATION DIVISION. PROGRAM-ID. REL02. ENVIRONMENT DIVISION. INPUT-OUTPUT SECTION. FILE-CONTROL. SELECT FLAVORS ASSIGN TO "BRAND" ORGANIZATION IS RELATIVE ACCESS MODE IS SEQUENTIAL. DATA DIVISION. FILE SECTION. FD FLAVORS. 01 KETCHUP-MASTER. 02 FILLER PIC X(14). 02 REC-NUM PIC 9(05). 02 FILLER PIC X(31). 02 FILLER PIC X(31). WORKING-STORAGE SECTION. 01 REC-COUNT PIC S9(5) VALUE 0. PROCEDURE DIVISION. A000-BEGIN. OPEN OUTPUT FLAVORS. PERFORM A010-WRITE 10 TIMES. CLOSE FLAVORS. STOP RUN. A010-WRITE. MOVE "Record number" TO KETCHUP-MASTER. ADD 1 TO REC-COUNT. MOVE REC-COUNT TO REC-NUM. WRITE KETCHUP-MASTER INVALID KEY DISPLAY "BAD WRITE" STOP RUN.

Creating a Relative File in Random Access Mode

When a program creates a relative file using random access mode, the program must place a value in the RELATIVE KEY data item before executing a WRITE statement. Example 6-25 shows how to supply the relative key. It writes 10 records in the cells numbered: 2, 4, 6, 8, 10, 12, 14, 16, 18, and 20. Record cells 1, 3, 5, 7, 9, 11, 13, 15, 17, and 19 are also created, but contain no valid records.

Example 6-25 Creating a Relative File in Random Access Mode

IDENTIFICATION DIVISION. PROGRAM-ID. REL03. ENVIRONMENT DIVISION. INPUT-OUTPUT SECTION. FILE-CONTROL. SELECT FLAVORS ASSIGN TO "BRAND" ORGANIZATION IS RELATIVE ACCESS MODE IS RANDOM RELATIVE KEY IS KETCHUP-MASTER-KEY. DATA DIVISION. FILE SECTION. FD FLAVORS. 01 KETCHUP-MASTER. 02 FILLER PIC X(14). 02 REC-NUM PIC 9(05). 02 FILLER PIC X(31). WORKING-STORAGE SECTION. 01 KETCHUP-MASTER-KEY PIC 99. 01 REC-COUNT PIC S9(5) VALUE 0. PROCEDURE DIVISION. A000-BEGIN. OPEN OUTPUT FLAVORS. MOVE 0 TO KETCHUP-MASTER-KEY. PERFORM A010-CREATE-RELATIVE-FILE 10 TIMES. DISPLAY "END OF JOB". CLOSE FLAVORS. STOP RUN. A010-CREATE-RELATIVE-FILE. ADD 2 TO KETCHUP-MASTER-KEY. MOVE "Record number" TO KETCHUP-MASTER. ADD 2 TO REC-COUNT. MOVE REC-COUNT TO REC-NUM. WRITE KETCHUP-MASTER INVALID KEY DISPLAY "BAD WRITE" STOP RUN.

Statements for Relative File Processing

Processing a relative file involves the following:

1. Opening the file
2. Setting the relative record number
3. Processing the file with valid I/O statements
4. Closing the file

Table 6-5 lists the valid I/O statements and illustrates the following relationships:

• Organization determines valid access modes.
• Organization and access mode determine valid open modes.
• All three (organization, access, and open mode) enable or disable I/O statements.

Table 6-5 Valid I/O Statements for Relative Files

			Open Mode
File Organization	Access Mode	Statement	INPUT	OUTPUT	I-O	EXTEND
RELATIVE	SEQUENTIAL	DELETE READ REWRITE START WRITE UNLOCK	No Yes No Yes No Yes	No No No No Yes Yes	Yes Yes Yes Yes No Yes	No No No No Yes Yes
	RANDOM	DELETE READ REWRITE WRITE UNLOCK	No Yes No No Yes	No No No Yes Yes	Yes Yes Yes Yes Yes	No No No No No
	DYNAMIC	DELETE READ READ NEXT REWRITE START WRITE UNLOCK	No Yes Yes No Yes No Yes	No No No No No Yes Yes	Yes Yes Yes Yes Yes Yes Yes	No No No No No No No

Writing a Relative File

Each WRITE statement places a record into a cell that contains no valid data. If the cell does not already exist, the I/O system creates it. To change the contents of a cell that already contains valid data, use the REWRITE statement.

6.3.4 File Handling for Indexed Files

Creating an indexed file involves the following tasks:

1. Specifying ORGANIZATION IS INDEXED in the Environment Division SELECT clause
2. Specifying ACCESS MODE IS SEQUENTIAL (or RANDOM or DYNAMIC) in the Environment Division SELECT clause
3. Opening the file for OUTPUT (to create and add records) or for I-O (to add, change, delete, or extend records)
4. Initializing the key values
5. Executing a WRITE statement
6. Closing the file

One way to populate an indexed file is to sequentially write the records in ascending order by primary key. Example 6-26 creates and populates an indexed file from a sequential file, which has been sorted in ascending sequence on the primary key field. Notice that the primary and alternate keys are initialized in ICE-CREAM-MASTER when the contents of the fields in INPUT-RECORD are read into ICE-CREAM-MASTER before the record is written.

Example 6-26 Creating and Populating an Indexed File

IDENTIFICATION DIVISION. PROGRAM-ID. INDEX02. ENVIRONMENT DIVISION. INPUT-OUTPUT SECTION. FILE-CONTROL. SELECT INPUT-FILE ASSIGN TO "DAIRYI". SELECT FLAVORS ASSIGN TO "DAIRY" ORGANIZATION IS INDEXED ACCESS MODE IS SEQUENTIAL RECORD KEY IS ICE-CREAM-MASTER-KEY ALTERNATE RECORD KEY IS ICE-CREAM-STORE-STATE WITH DUPLICATES ALTERNATE RECORD KEY IS ICE-CREAM-STORE-CODE. DATA DIVISION. FILE SECTION. FD INPUT-FILE. 01 INPUT-RECORD. 02 INPUT-RECORD-KEY PIC 9999. 02 INPUT-RECORD-DATA PIC X(47). FD FLAVORS. 01 ICE-CREAM-MASTER. 02 ICE-CREAM-MASTER-KEY PIC XXXX. 02 ICE-CREAM-MASTER-DATA. 03 ICE-CREAM-STORE-CODE PIC XXXXX. 03 ICE-CREAM-STORE-ADDRESS PIC X(20). 03 ICE-CREAM-STORE-CITY PIC X(20). 03 ICE-CREAM-STORE-STATE PIC XX. WORKING-STORAGE SECTION. 01 END-OF-FILE PIC X. PROCEDURE DIVISION. A000-BEGIN. OPEN INPUT INPUT-FILE. OPEN OUTPUT FLAVORS. A010-POPULATE. PERFORM A100-READ-INPUT UNTIL END-OF-FILE = "Y". A020-EOJ. DISPLAY "END OF JOB". STOP RUN. A100-READ-INPUT. READ INPUT-FILE INTO ICE-CREAM-MASTER AT END MOVE "Y" TO END-OF-FILE. IF END-OF-FILE NOT = "Y" WRITE ICE-CREAM-MASTER INVALID KEY DISPLAY "BAD WRITE" STOP RUN.

The program can add records to the file until it reaches the physical limitations of its storage device. When this occurs, you should follow these steps:

1. Delete unnecessary records.
2. Back up the file.
3. Recreate the file either by using the OpenVMS Alpha CONVERT Utility to optimize file space, or by using a Compaq COBOL program.

Statements for Indexed File Processing

Processing an indexed file involves the following:

1. Opening the file
2. Processing the file with valid I/O statements
3. Closing the file

Table 6-6 lists the valid I/O statements and illustrates the following relationships:

• File organization determines valid access modes.
• File organization and access mode determine valid open modes.
• All three (organization, access, and open mode) enable or disable I/O statements.

Table 6-6 Valid I/O Statements for Indexed Files

			Open Mode
File Organization	Access Mode	Statement	INPUT	OUTPUT	I-O	EXTEND
INDEXED	SEQUENTIAL	DELETE READ REWRITE START WRITE UNLOCK	No Yes No Yes No Yes	No No No No Yes Yes	Yes Yes Yes Yes No Yes	No No No No Yes Yes
	RANDOM	DELETE READ REWRITE WRITE UNLOCK	No Yes No No Yes	No No No Yes Yes	Yes Yes Yes Yes Yes	No No No No No
	DYNAMIC	DELETE READ READ NEXT REWRITE START WRITE UNLOCK	No Yes Yes No Yes No Yes	No No No No No Yes Yes	Yes Yes Yes Yes Yes Yes Yes	No No No No No No No

Writing an Indexed File

You specify sequential access mode in the Environment Division SELECT clause when you want to write records in ascending or descending order by primary key, depending on the sort order. Specify random or dynamic access mode to enable your program to write records in any order.

Using Segmented Keys in Indexed Files

Segmented keys are a form of primary or alternate keys. A segmented key can be made up of multiple pieces, or segments. These segments are data items that you define in the record description entry for a file. They are concatenated, in order of specification in the ALTERNATE RECORD KEY or RECORD KEY clause, to form the segmented key, which will be treated like any "simple" primary or alternate key.

With segmented keys, you have more flexibility in defining record description entries for indexed files. A segmented key is made up of between one and eight data items, which can be defined anywhere and in any order within the record description, and which can even overlap. For example, you might use the following record definition in your program:

01 EMPLOYEE. 02 FORENAME PIC X(10). 02 BADGE-NO PIC X(6). 02 DEPT PIC X(2). 02 SURNAME PIC X(20). 02 INITIAL PIC X(1).

Then the following line in your program, which specifies the segmented key name and three of its segments:

RECORD KEY IS NAME = SURNAME FORENAME INITIAL

causes Compaq COBOL to treat name as if it were an explicitly defined group item consisting of the following:

02 SURNAME PIC X(20). 02 FORENAME PIC X(10). 02 INITIAL PIC X(1).

You define a segmented key in either the RECORD KEY clause or the ALTERNATE RECORD KEY clause. You use the START or READ statement to reference a segmented key.

Each segment is a data-name of a data item in a record description entry. A segment can be an alphanumeric or alphabetic item, a group item, or an unsigned numeric display item. A segment can be qualified, but it cannot be a group item containing a variable-occurrence item.

See the chapters on the Data Division and the Procedure Division in the Compaq COBOL Reference Manual for more information on segmented keys.

Example 6-27 shows how you might use segmented keys. In this example, SEG-ICE-CREAM-KEY is a segmented-key name. ICE-CREAM-STORE-KIND and ICE-CREAM-STORE-ZIP are the segments. Notice that the segmented-key name is referenced in the READ statement.

Example 6-27 Using Segmented Keys

IDENTIFICATION DIVISION. PROGRAM-ID. MANAGER. ENVIRONMENT DIVISION. INPUT-OUTPUT SECTION. FILE-CONTROL. SELECT FLAVORS ASSIGN TO "STORE" ORGANIZATION IS INDEXED ACCESS MODE IS RANDOM RECORD KEY IS SEG-ICE-CREAM-KEY = ICE-CREAM-STORE-KIND, ICE-CREAM-STORE-ZIP. DATA DIVISION. FILE SECTION. FD FLAVORS. 01 ICE-CREAM-MASTER. 02 ICE-CREAM-DATA. 03 ICE-CREAM-STORE-KIND PIC XX. 03 ICE-CREAM-STORE-MANAGER PIC X(40). 03 ICE-CREAM-STORE-SIZE PIC XX. 03 ICE-CREAM-STORE-ADDRESS PIC X(20). 03 ICE-CREAM-STORE-CITY PIC X(20). 03 ICE-CREAM-STORE-STATE PIC XX. 03 ICE-CREAM-STORE-ZIP PIC XXXXX. WORKING-STORAGE SECTION. 01 PROGRAM-STAT PIC X. 88 OPERATOR-STOPS-IT VALUE "1". PROCEDURE DIVISION. A000-BEGIN. OPEN I-O FLAVORS. PERFORM A020-INITIAL-PROMPT. IF OPERATOR-STOPS-IT PERFORM A005-TERMINATE. PERFORM A030-RANDOM-READ. PERFORM A025-SUBSEQUENT-PROMPTS UNTIL OPERATOR-STOPS-IT. PERFORM A005-TERMINATE. A005-TERMINATE. DISPLAY "END OF JOB". STOP RUN. A020-INITIAL-PROMPT. DISPLAY "Do you want to see the manager of a store?". PERFORM A040-GET-ANS UNTIL PROGRAM-STAT = "Y" OR "y" OR "N" OR "n". IF PROGRAM-STAT = "N" OR "n" THEN MOVE "1" TO PROGRAM-STAT. A025-SUBSEQUENT-PROMPTS. MOVE SPACE TO PROGRAM-STAT. DISPLAY "Do you want to see the manager of another store?". PERFORM A040-GET-ANS UNTIL PROGRAM-STAT = "Y" OR "y" OR "N" OR "n". IF PROGRAM-STAT = "Y" OR "y" THEN PERFORM A030-RANDOM-READ ELSE MOVE "1" TO PROGRAM-STAT. A030-RANDOM-READ. DISPLAY "Enter store kind: ". ACCEPT ICE-CREAM-STORE-KIND. DISPLAY "Enter zip code: " AT LINE PLUS 2. ACCEPT ICE-CREAM-STORE-ZIP. PERFORM A100-READ-INPUT-BY-KEY. A040-GET-ANS. DISPLAY "Please answer Y or N" ACCEPT PROGRAM-STAT. A100-READ-INPUT-BY-KEY. READ FLAVORS KEY IS SEG-ICE-CREAM-KEY INVALID KEY DISPLAY "Store does not exist - Try again" NOT INVALID KEY DISPLAY "The manager is: ", ICE-CREAM-STORE-MANAGER.

6.4 Reading Files

Reading sequential, line sequential, relative, and indexed files includes the following tasks:

1. Opening the file
2. Executing a READ or START statement

Sections 6.4.1, 6.4.2, and 6.4.3 describe the specific tasks involved in reading sequential, line sequential, relative, and indexed files.

6.4.1 Reading a Sequential or Line Sequential File

Reading a sequential or line sequential file involves the following:

1. Opening the file for INPUT or I/O for sequential files, or INPUT for line sequential files (I/O is not permitted for line sequential files)
2. Executing a READ statement

Each READ statement reads a single logical record and makes its contents available to the program in the record area. There are two ways of reading records:

• READ file-name INTO destination-area
• READ file-name

Statements (1) and (2) in the following example are logically equivalent:

FILE SECTION. FD STOCK-FILE. 01 STOCK-RECORD PIC X(80). WORKING-STORAGE SECTION. 01 STOCK-WORK PIC X(80). -------------(1)--------------- -------------(2)--------------- READ STOCK-FILE INTO STOCK-WORK. READ STOCK-FILE. MOVE STOCK-RECORD TO STOCK-WORK.

When you omit the INTO phrase, you process the records directly in the record area or buffer (for example, STOCK-RECORD). The record is also available in the record area if you use the INTO phrase.

In a READ INTO clause, if the destination area is shorter than the length of the record area being read, the record is truncated on the right and a warning is issued; if longer, the destination area is filled on the right with blanks.

If the data in the record being read is shorter than the length of the record (for example, a variable-length record), the contents of the record beyond that data are undefined.

Generally speaking, if the recordtype is fixed, the prolog and epilog are zero. The exceptions to this are: for relative files there is a 1 byte record status flag prolog; for sequential files there is a 1 byte epilog if the record length is odd.

Example 6-28 reads a sequential file and displays its contents on the terminal.

Example 6-28 Reading a Sequential File

IDENTIFICATION DIVISION. PROGRAM-ID. SEQ02. ENVIRONMENT DIVISION. INPUT-OUTPUT SECTION. FILE-CONTROL. SELECT TRANS-FILE ASSIGN TO "TRANS". DATA DIVISION. FILE SECTION. FD TRANS-FILE. 01 TRANSACTION-RECORD PIC X(25). PROCEDURE DIVISION. A000-BEGIN. OPEN INPUT TRANS-FILE. PERFORM A100-READ-TRANS-FILE UNTIL TRANSACTION-RECORD = "END". CLOSE TRANS-FILE. STOP RUN. A100-READ-TRANS-FILE. READ TRANS-FILE AT END MOVE "END" TO TRANSACTION-RECORD. IF TRANSACTION-RECORD NOT = "END" DISPLAY TRANSACTION-RECORD.