Re: Unprint(display)able chars in Data files

[CRPence <CRPbottle@xxxxxxxxx>]

  So ASCII or UTF data is being stored in an EBCDIC column, and the 
desire is to suppress the code points in the data which are not valid 
displayable\printable EBCDIC characters?  And what if the code point is 
presented as a glyph that can be understood by the user, but for which 
there is no appropriate meaning for its presence?  Why not just store 
the data in a column that reflects what the data really is, or convert 
the given data into the CCSID of the column; i.e. do not lie to the 
database or to the programs, about what the data is, in that field?

  If the data will not be represented properly and thus data from thos 
column(s) will need to be translated when the data is read by a program, 
then deferring the work to the database is possible via a logical file 
using the Translation Table (TRNTBL) DDS field-level keyword.  The xlate 
table can be created using CRTTBL /* SRCFILE(*PROMPT) TBLTYPE(*CVT) 
BASETBL(*HEX) */.

Regards, Chuck

On 05 Jun 2012 02:14, Gad Miron wrote:
> The uploaded files are transactions reported by our customers using
> their items (SKUs and descriptions.)
>
> We then map those SKUs to our own Catalog Numbers, hence the need to
> keep the external text descriptions with all those unprintable
> characters. The data are written to "simple" DB files that is to say
> CCSID 424 (not ASCII or UTF).
>
> Since there are several text data fields in each record I would like
> to use a single XLATE for each text field.
>
> I could define a procedure to loop through the text and to replace
> and invoke it for each field but EXLATE seems more elegant to me ...