Introduction to Sybase, Part 2

This month Mr. Sissom shows us how to set up and use a Sybase client written in Perl through examples.
Writing a sybperl Program

All of the programs here are available for download. If you type in these programs, be sure to use chmod to make them executable.

Listing 1.

Writing a sybperl program is quite simple. Listing 1 is our first example program. This will list the names of all the databases in the server. Here is a line-by-line explanation of the program.

Line 1 tells Linux which program to use to run this script. This must be the new version of Perl you just installed. Make sure you change this line to point to the correct version of Perl on your system.

Line 3 tells Perl to use the CT-library interface to Sybase. It should be at the beginning of all Perl scripts you write that access a Sybase server.

Line 5 attaches to the correct Sybase server. The first parameter is the user name, the second is the password and the third is the name of the server.

Line 7 is the SQL to run.

Lines 9-10 are commands that run the query on the server and return a reference to an array of rows, @rows. Note this command loads the entire result set into memory. This is fine for small result sets, but if you are expecting a large result set, you shouldn't use ct_sql. Later, I will give you an alternative method for executing commands and receiving large result sets.

Lines 12-14 will print all rows that were returned.

Listing 2

Listing 2 is an example of a sybperl program that updates data. In Line 7, we use the same ct_sql command to send the SQL to the database, except this time a set of rows is not returned. The insert, delete and update SQL commands also do not return rows. The SQL command use pubs2 tells Sybase to make the pubs2 database the default database for the rest of this session. In Line 10, we again use the ct_sql command to run the SQL. This time, we add a row to the discounts table. You can use the isql program to run an SQL SELECT command to verify that the row was added.

Writing a Perl CGI Client Program

Linux is mostly used as a web server, and Perl is primarily used to write web applications. So, we will create a Perl program to access the Sybase database.

Listing 3

Writing a CGI program to access Sybase is quite simple. Listing 3 is the complete code of a CGI program to let you know who's logged in to your Sybase server. Place this program in your web server's cgi-bin directory. On a default Red Hat system, the directory is /home/httpd/cgi-bin/. For this example, name the program

In lines 5 and 6, we set two environment variables. The Sybase DB-Library and CT-Library must find these environment variables, or an error will occur. When you run a CGI program, very few environment variables are passed to your program. These two environment variables must be set in each CGI program that needs to access Sybase. If you have many CGI programs, place these commands in a file included in all your CGI programs.

The SYBASE environment variable contains the directory of the Sybase software. The DSQUERY variable contains the name of the default server.

The only other difference between this example and the others is it outputs HTML to a browser.

Other Considerations

These example programs show the basics of accessing a Sybase database server. In production programs, a few more things must be taken care of in your programs.

Errors from the server must be handled properly. If you ignore them, your program will stop when it encounters a server error.

In all our example programs, ct_sql was used to run queries. It works fine for SQL commands and stored procedures which don't return result sets, but would have severe problems for queries returning large result sets.

Listing 4

Listing 4 shows how to handle errors and demonstrate a replacement for the ct_sql command. In lines 3 and 4, we establish both a client and a server message call-back routine. These routines will be called when the server or client generates an informational or error message.

In lines 7-20, we handle a single SQL statement. Sybase allows a single statement to return multiple result sets. Lines 8-10 will process each result set. Lines 17-19 will handle each row in a result set. Lines 11-16 will look at the result set and print the names and types of each column. A Sybase result set contains more than just data—it also includes column definition information.

Lines 23-50 are the two call-back routines. These routines are called each time there is a message from the server or client. An example of a client message is the one returned if you can't log in to the server. An example of a server message is the one returned if you have an error in your SQL.

All this information can be found within the sybperl and the Sybase documentation.