Making the SQL driver available to PERCobol:
The JDBC (Java DataBase Connectivity) driver must be available to therunning program. The
JDBC driver is specific to the database, in this case SQL, and these JDBC drivers come
from the database vendor such as IBM or Oracle. It will be a file ending in .zip or .jar
(do not extract it). The filename for it changes over time, so check your database
release's documentation.
To make the driver available to a PERCobol project within the IDE, right-click the
project, select Properties. This is where the Cobol and Java settings for the project are
available. Since this is an element of Java functionality being exposed to PERCobol,
select Java Build Path. The libraries tab lists all Java libraries available to the
program. Using the 'External JARs' button, add the SQL JDBC driver library to the
project. It's now available to the program at runtime.
While in the properties, select Cobol Compiler Settings from the left.
Select Category SQL. In the JDBC tab is Enable enhancements
for SQL SQL/JDBC; check it.
Making the SQL connection:
To enable it from the program, you must connect to the database before you may use it. The
typical connection statement is of the form:
EXEC SQL
CONNECT TO jdbc:SQL://host:6789/dbname
USER :uid
PASSWORD :pwd
DRIVER COM.ibm.SQL.jdbc.net.SQLDriver
SERIALIZABLE
END-EXEC
The CONNECT TO piece is the URL for the connection as
documented in the SQL JDBC manual. It always begins with jdbc:. The SQL: specifies which
driver to use. The host is the hostname location of the database. 6789 is the default port
for SQL, but if you have SQL using another port change it. And dbname is specific to your
setup.
:uid and :pwd are just Cobol host variables. They may be specified
directly if desired, although the user is often prompted for password so host variables
are most often used.
The DRIVER specifies the classname of the JDBC driver. The classname is given in
the JDBC driver's documentation, usually by showing a snippet of Java code
Class.forName("drivername"); JDBC is often not documented well. The driver
listed above is often the most common SQL remote driver name. (SQL local to the machine
has a faster driver with a separate name; there are often two included with SQL, one
remote and one local.)
SERIALIZABLE is an option, but used quite often.
The connection to SQL from within an application server will be a little different, but
all changes are encapsulated within the connect
statement. If running in a managed environment such as an application server where it says
a DataSource is available, the CONNECT TO piece is just ds:name where name is
exported by the application server and the driver clause is not required.
PERCobol SQL general:
SQLCODE and SQLSTATE can be any definition sufficient to hold the data, such as PIC S999
and PIC X(5). Their usage clauses don't matter.
When creating a new program, you can select SQL Program as a template to give a basic
reminder of the SQL uses.
|
 |
