On Thu04-Apr-2011 16:27 , James Lampert wrote:
Names delimited with quotation marks will maintain mixed-case. Of
course lower-case alphabetic characters are variant rather than
invariant . . .
Not in EBCDIC, they're not!
In EBCDIC, the entire upper and lower case alphabet, all 10 digits,
space, &, -, /, :, period, comma, (, ), _, single and double
typewriter quotes, +, ?, =, ;, and > are invariant, and < is
One could wish that lowercase characters were invariant, but
unfortunately to suggest they are, is not entirely accurate.
Globalization documentation for thy IBM i OS specifically warns, among
other things, that "Lowercase alphabets should not be assumed to be
invariant." Consider that the EBCDIC code page 290 either has no
lowercase a-z or those characters are variant. Perform the following
scripted requests on a IBM i command line, honoring the specified case,
to infer there is an issue:
ChgJob CCSID(37) /* Fails: "Command ChgJob ... not found" */
If no CCSID based on code page 290 is going to be supported, then I
suppose lower case characters from the character set 00640 used in SQL
identifiers should be fine.
I believe both the "required space" and "quotation mark" characters
are quasi-invariant [he-heh; OK, rarely variant] for being invariant
across all but one or two EBCDIC CCSIDs, however I am unsure of the
accuracy of that claim about the less-than character. I do not recall
ever testing SQL for the variant delimiter, the [double] quote being a
possible problem. Perhaps the '<' character might be missing from one
or more default keyboard layouts, but I do not recall that character is
variant in any EBCDIC CCSID; though perhaps the Euro character changed
something I do not recall.?