Several of our products include the ability to convert data from one encoding
to another. Typically source and target Coded Character Set Identifiers
(CCSIDs) are specified on the command performing the product's function.
Our products access the extraordinary national language support (NLS) capabilities
of i5/OS
through calls to the iconv() application programming interface (API) which
converts data between CCSIDs. Correct CCSID selection is crucial to accurate
conversions results. Extensive information to help you select CCSIDs is
available in IBM's®Information Center,
search on "Globalization reference information" to locate it.
But before you dig in, we might be able to save you some time. On the vast
majority of single language systems, the CCSID that data is encoded in is
strongly influenced by the keyboard type (system value QKBDTYPE) used by
the system. And in turn this source CCSID strongly influences the useful,
usually ASCII, target CCSIDs of interest. For most single language systems
you can find the source CCSID directly by running DSPJOB
OPTION(*DFNA) and locating the "Default coded character set
identifier" value. Once you have a source CCSID the table below offers
suggestions on usually useful target CCSIDs to try.
The final test for the accuracy of your CCSID selections is whether or
not your data has been properly converted. Check your results.
A special note for PDF
conversions. By default, our products' PDF
conversions utilize Acrobat's built in fonts. This reduces file size. Acrobat's
built in fonts have a limited number of character sets and therefore this
default does not always give usable results. If your source data contains
characters not found in Acrobat's character sets, the built in fonts will
not work for you. If Acrobat's limitation affects you, the way around it
is to imbed the fonts containing characters unknown to Acrobat in the generated
PDF.
Imbedding can be accomplished in either of two ways. One way is to specify
PDF
font imbedding on the product's command. The other is to specify a PDF/A
transform which automatically imbeds fonts. Roughly speaking, if your source
CCSID doesn't convert to target CCSID 819 (*ISO88591) or isn't Chinese,
Japanese, or Korean, you must imbed fonts to get useful results. Also note,
for Chinese, Japanese, or Korean specify a target CCSID of *UNICODE for
best results.
Suggestions
For a given source Coded Character Set Identifier (CCSID) this table suggests
typically useful, often ASCII, target CCSIDs that may give the results you
are looking for. In all cases, check the output for accuracy. Predominately,
the target CCSIDs shown are the 4100 and 4105 encoding scheme "associated
CCSID" returned by i5/OS's
Coded Data Representation Architecture (CDRA) Get Related Default CCSID
(CDRGRDC) API.
If you know useful suggestions that should be included here, please send
us a note at
.
i5/OS's
QKBDTYPE (keyboard type) system value often implies a related EBCDIC CCSID.
The following table shows this relationship. You can determine your system's
setting by running DSPSYSVAL QKBDTYPE.
