-
Result will be different in a UNICODE or ASCII system vs an EBCDIC system, why didn't you try with a simple report (check online documentation for such program, look for 'Executable Examples')
-
Raymond Giuseppi as far as I can see in the documentation, with Unicode systems, SORT without AS TEXT doesn't depend anymore on the code page but only on the Unicode code point, and SORT with AS TEXT depends on the locale (based on the logon language or the one set by SET LOCALE).
Debanjan Das Concerning SORT with AS TEXT, for example the locales for languages Serbian, Chinese and Thaï, have characters sequenced as follows, here from the 4900th position; both the glyph and its Unicode code point are represented; it's done by a custom program:
SAP online example gives this result (1st sorting is based on the Unicode code points, 2nd and 3rd on the order of characters in the current locale):
-
so in general, the alphanumeric would be understood by computer in term of sequence of 2 bit 0 & 1. then sort is compare those sequence.
2 Answers
-
Posted on Jul 01, 2019 at 06:46 AM
Character-like columns are sorted by default by their binary representation, encoding in the code page of the current text environment.
You can force a 'textual' sort using option AS TEXT to use locale text environment, which depends on Language, Operating System and Kernel (e.g. rules fo upper/lower case or accented characters, or characters that don't belong to the language...) Can you guess performance impact of this option...
Add a comment
-
Posted on Jul 01, 2019 at 07:51 AM
It will sort
- ABC1234
- ABC7865
- ABC0859
- ABB9999
- 90
- ANHYT
alphanumerically. It seems you don't know what "alphanumerically" means, so I suggest you write a short program and find out exactly how this data will be sorted on your system..
Add a comment
Add a comment