Tweet
COPY...REPLACING in Cobol's SELECT Statement
One of Cobol's great strengths is its ability to replace one
string in a source member with another at compile time.
This feature lets Cobol programmers create generic,
pseudo- parametized "template" source members that can be
used in a variety of different cases.
For instance, you could declare a database file to a Cobol
program in the ENVIRONMENT DIVISION using the following
SELECT statement:
If, instead of hard-coding the SELECT statement,
you preferred to create a generic source member to
copy, you might try to code
Unfortunately, it doesn't work; the Cobol compiler chokes
when it reaches DATABASE- os400name because COPY...
REPLACING can replace only entire tokens, and os400name
is not a token. DATABASE-os400name is the token.
There are two solutions. The first one is to code the
COPY...REPLACING statement to replace the entire token;
in other words, to replace DATABASE-os400name with
DATABASE- custmast. But that's clumsy.
It's much better to write the SELECT statement as follows,
with parentheses around the part of the token you want to replace:
Now you can code the COPY...REPLACING statement like this:
This method will give you the result you want. The technique takes
advantage of the fact that the open parenthesis is a valid terminator
for a Cobol word. The parentheses cause the Cobol compiler to break
DATABASE-(os400name) into four tokens: DATABASE-, (, os400name, and ).
This technique produces compilable source code even though SEU's
built-in syntax checker rejects the parentheses. To save yourself
time and prevent frustration, turn off syntax checking by pressing
the F13 key and typing an N (no) at the appropriate input field.
Jon Paris
One of Cobol's great strengths is its ability to replace one
string in a source member with another at compile time.
This feature lets Cobol programmers create generic,
pseudo- parametized "template" source members that can be
used in a variety of different cases.
For instance, you could declare a database file to a Cobol
program in the ENVIRONMENT DIVISION using the following
SELECT statement:
Code:
SELECT file-name
ASSIGN TO DATABASE-os400name
ORGANIZATION IS org-type
ACCESS MODE IS access-type.
you preferred to create a generic source member to
copy, you might try to code
Code:
COPY selectx REPLACING
==file-name== BY ==cust-file==
==os400name== BY ==custmast==
==org-type== BY ==INDEXED==
==access-type== BY ==SEQUENTIAL==.
to obtain
SELECT cust-file
ASSIGN TO DATABASE-custmast
ORGANIZATION IS INDEXED
ACCESS MODE IS SEQUENTIAL.
when it reaches DATABASE- os400name because COPY...
REPLACING can replace only entire tokens, and os400name
is not a token. DATABASE-os400name is the token.
There are two solutions. The first one is to code the
COPY...REPLACING statement to replace the entire token;
in other words, to replace DATABASE-os400name with
DATABASE- custmast. But that's clumsy.
It's much better to write the SELECT statement as follows,
with parentheses around the part of the token you want to replace:
Code:
SELECT file-name
ASSIGN TO DATABASE-(os400name)
ORGANIZATION IS org-type
ACCESS MODE IS access-type.
Code:
COPY selectx REPLACING
==file-name== BY ==cust-file==
==(os400name)== BY ==custmast==
==org-type== BY ==INDEXED==
==access-type== BY ==SEQUENTIAL==.
advantage of the fact that the open parenthesis is a valid terminator
for a Cobol word. The parentheses cause the Cobol compiler to break
DATABASE-(os400name) into four tokens: DATABASE-, (, os400name, and ).
This technique produces compilable source code even though SEU's
built-in syntax checker rejects the parentheses. To save yourself
time and prevent frustration, turn off syntax checking by pressing
the F13 key and typing an N (no) at the appropriate input field.
Jon Paris