This database question should be asked of a Database Administrator (DBA) with advanced mainframe experience to show the candidate's understanding of the use of ASCII, EBCDIC and UNICODE code pages. A DBA with 3-4 years experience who has kept up-to-date in his or her knowledge should be able to express a solution to this problem fairly quickly.

We have an existing application that provides information for our customers throughout the United States who call into a customer care center. We'd like to do two things. The first is to expand the application to our European customers. The second is to make the application accessible via the Web.

Currently, the database uses an EBCDIC code page. Are there any issues moving forward using this code page with the different European currencies, including the Euro? What other things do we need to be aware of as we roll out this application via the Web?

A good candidate will explain that there are may different EBCDIC code pages, a code page being a method that assigns unique bit patterns to characters. Within DB2 there is the ability to change the CCSID from one that doesn't support the Euro symbol to one that does. The CCSID tells DB2 what code page to use. A good candidate will also note that the process to change CCSIDs is not a simple matter. The data will need to be unloaded, the table space definition altered and the data reloaded using a LOAD utility.

The good candidate will also follow up on the second question, noting that non-mainframe servers generally use ASCII code pages. He or she would also know that the EBCDIC and ASCII encoding schemes have different sort orders. This needs to be taken into account if the mainframe database is going to stay in EBCDIC, but be used by people who are used to seeing things in an ASCII sort order.

An excellent candidate would also know that the EBCDIC data is translated into ASCII data by DB2 Connect when it is retrieved from the mainframe. If ASCII data is sent to the mainframe, it is translated to EBCDIC at the mainframe. An excellent candidate might also mention that this translation is not always perfect, especially when special characters are involved and that thorough testing is a good way to determine if any of these problems exist.

A follow up question might be "How do I avoid the extra overhead of conversion and a different sort order?" Most candidates will know that data can be stored on the mainframe in ASCII format.

A final question might be, "I've heard about UNICODE. What is that?"

If the candidate tells you that UNICODE will translate data from one language to another (don't laugh -- some people believe this), he or she is not a very knowledgeable candidate. A good candidate will tell you that UNICODE is another encoding scheme that contains information for the majority of international characters, including double-byte character sets such as those used in Japan. An excellent candidate will tell you that there are two main variations on UNICODE -- UTF-8 and UTF-16, the numbers telling you the number of bits making up the encoding scheme.

About the author

Casey Young is an eclectic writer whose work spans poetry, children's writing, fiction and technical writing. She has worked with DB2 for over seventeen years, and is currently manager of the DB2 performance group at IBM's Silicon Valley Lab. She has worked on designing and tuning some of the world's largest databases and is author of "Exploring IBM e-business Software" available at www.maxpress.com.