Question

I just can’t find my mistake…

I have a csv-file with lines like this:

11;1116209173900;8;4690;000;ÖBB;20090831;20100330;O603;603-Deutschlandsberg;

I’d like to remove the O in front of 603.
My sed command looks like this:

sed 's/\(\(.*;\)\{8\}\)O\(.\{3\}\);/\1\2;/g'

but what I get is

11;1116209173900;8;4690;000;ÖBB;20090831;20100330;20100330;;603-Deutschlandsberg;

Where’s my mistake?

Thank you in advance!

Answer 1

Just a very small change:

sed 's/\(\(.*;\)\{8\}\)O\(.\{3\}\);/\1\3;/'

You have two levels of parentheses (\(\(.*;\)\{8\}\)), i.e. \2 contains the last match of .*;.
Btw.: For performance and unambiguousness reasons I think it’s better to use [^;]* instead of .*:

sed 's/\(\([^;]*;\)\{8\}\)O\(.\{3\}\);/\1\3;/'