Arabic is a complex script that is written from right to left. The characters have four basic forms according to the position that they take within the word: initial, middle, final, or isolated. Also, it uses ligatures, which are characters that consist of two or more characters. Most notably, the lam-alef ligature consists of lam (initial or medial form) with alef (final form).

In some encodings, such as EBCDIC420, the lam-alef ligature has a single code point whereas others, such as Windows 1256 or ISO-8859-6, encode the lam and alef characters as separate code points. When SAS transcodes data from EBCDIC420 to one of the other Arabic single-byte encodings, the lam-alef combined character cannot be mapped to a single code point, so a substitution character is placed in the data instead.

A new function, ANORM420, is being introduced in SAS^® 9.4 TS1M1. It correctly maps the EBCDIC 420 lam-alef to two separate code points in your data. The function also has options that enable you to add spaces after code points representing the final form of a character and convert Arabic-Indic numerals to a digit.

The following example shows the ANORM420 function usage.

s1=C8E5C7E3C7D320A02020

s2=C8E5C7E3C7D3A0202020

s3=59CD57BC577740454040

Click the Hot Fix tab in this note to access the hot fix for this issue.

Operating System and Release Information

A fix for this issue for Base SAS 9.4 is available at: