Facts and Myths

 

ODBC myths are erroneous information that is usually a byproduct of not reading the documentation or not actually using ODBC. Often a persistent rumor is in direct contradiction with the ODBC 3.0 Programmers's Reference and SDK Guide. Other times the behavior attributed to ODBC is actually related to the application, database engine, or object layer that uses ODBC. Here are some of the most common ODBC myths:

Myth: ODBC is not scaleable technology suitable for large systems.

• The truth is that ODBC is used with very large databases, for applications processing a large volume of daily transactions, and for applications having large number of users. Examples include Wal*Mart (5 terabyte database), MicroBilt (2.5 gigabytes per day) and Xerox PeopleNet (15-20,000 users).

Myth: ODBC is an SQL API for Windows programs.

• ODBC has been multi-platform for several years with versions for Unix, Macintosh, OS/2 and Windows operating systems. Microsoft ODBC 3.0 is an implementation of the ANSI/ISO SQL-92 Call Level Interface, the most recent international standard for SQL. The ODBC SDK for Windows platforms is available from Microsoft. For other platforms it is available from Intersolv.

Myth: ODBC supports only numbers and characters, not BLOBs or CLOBs.

• ODBC's SQLGetData and SQLPutData functions support large binary and character objects.

Myth: You can't program ODBC applications using Java.

• Javera's provides classes (JETConnect) that work with ODBC and JDBC drivers. Many middleware products support an ODBC connection to the database, and a JDBC connection to Java clients (type-3 JDBC drivers).

Myth: ODBC is slow.

• ODBC drivers now deliver equivalent or better performance than proprietary APIs. Consider that published benchmarks and articles refute the myth of poor ODBC performance. For more information read the ODBC Performance page by clicking here.

copyright (C) 1997, Ken North

ODBC Vendor Information
Ken North Home Page