I have the line of text within a text file. The line looks something like this:
xxxx,xxxxx,xxxxxx,xxxxx,xxxx,NL-1111 xx,xxxx,xxx
The NL- is an identifier for the country so this could be anything. I would like to remove the NL- part from the line so it looks like this:
xxxx,xxxxx,xxxxxx,xxxxx,xxxx,1111 xx,xxxx,xxx
And write the file afterwards.
Thanks in advance.
Another solution close to sed's ones, but with perl:
perl -i -pe "s/(?<=,)[a-zA-Z]{2}-//g" file.txt
It uses look behind expression, so that you don't need to repeat the comma in the replacement part.
sed -i); see my answer for details.
-i: on some (on my Windows cmd for example), you need to specify a backup pattern after -i (no backup is not possible). With that, it works everywhere, but you get a backup file.something like this using sed
sed -i 's/,[A-Z][A-Z]-\([0-9]\+,\)/,\1/i' file.txt
,[A-Z][A-Z]-\([0-9]\+,\)search for comma letter, letter, -, digit(s), comma
,\1keep only the commas and the digits.
iignore case on the letters
thankyou to @chris for proof-reading.
< necessary?I think the simplest solution here is reading it from the file into a shell variable, then writing it back immediately and using the pattern substitution variation of parameter expansion:
line="$(<file)"; echo "${line/[a-zA-Z][a-zA-Z]-}" >|file;
I would warn you against solutions that use sed-in-place functionality. I've found that sed behavior differs on different platforms with respect to the -i option. On Mac you have to give an empty argument ('') to the -i option, while on Cygwin you must not have an empty argument following the -i. To get platform compatibility you'd have to test what platform you're on.
sed might do the trick: remove the string ",NL-", "BE-" etc from anywhere in the file:
sed -i 's/,[A-Z][A-Z]-/,/' file.txt
