[SOLVED] sed substitute one of two strings

keirvt · 10-01-2015, 04:01 PM

I want to use sed to remove a string with capitals that is followed by a comma, then space or just a space then a country name that begins with a captital. I achieve this at present with this pair of sed commands;

Code:

sed 's/[A-Z]\+ \([A-Z]\)/\1/'|sed 's/[A-Z]\+, \([A-Z]\)/\1/'

So for example, a file, "infile" with the lines
|
UCTEXT, Australia
MOREUCTEXT Zambia

will output

Australia
Zambia

Is there a way to do this with a single sed regex? It would be nice if something like below existed

Code:

sed 's/[A-Z]\+[" ",", "] \([A-Z]\)/\1/'

syg00 · 10-01-2015, 06:17 PM

Use extended regex (-r) and use (round) parentheses for the "or" - ( |, )
Note this will change the back-reference you want to use.
As a matter of style, perhaps better to use character classes to make the regex codepage safe - [[:upper:]] rather than [A-Z]