In the case of simple databases, such as flat-file or Berkeley DB files, ``connecting'' is usually as simple as opening the files for reading or using the tie mechanism. However, in larger database systems, connecting may be considerably more complicated.

A relatively simple RDBMS is mSQL, which has a simple method of connection: to connect, a program connects to a TCP/IP port on the computer running the database. This establishes a live connection within the database. However, more complex systems, such as Oracle, have a lot more internal security and housekeeping work that must be performed at connection time. They also have more data that needs to be specified by the program, such as the username and password that you wish to connect with.

By looking at a broad spectrum of database systems, the information required to connect can be boiled down to:

  1. The data source name, a string containing information specifying the driver to use, what database you wish to connect to, and possibly its whereabouts. This argument takes the format discussed in the previous section and is highly database-specific.

  2. The username that you wish to connect to the database as. To elaborate on the concept of usernames a little further, some databases partition the database into separate areas, called schemas, in which different users may create tables and manipulate data. Users cannot affect tables and data created by other users. This setup is similar to accounts on a multiuser computer system, ...

Get Programming the Perl DBI now with the O’Reilly learning platform.

O’Reilly members experience live online training, plus books, videos, and digital content from nearly 200 publishers.