How can I properly use the 'sed' s command in this particular scenario

Question

I have a names.txt file where each line is of the form:

xxxxxx   random_string_of_characters    2015

where xxxxxx is a 6-digit number, and random_string_of_characters can be anything. I want to use the substitute command to replace all the empty space and the random_string_of_characters between xxxxxx and 2015 in each row, so that each string looks like this:

xxxxxx 2015

So, what would be the best way to accomplish this?

Answer 1

You could do

sed -i -e 's/[[:space:]]\+.\+2015$/ 2015/' names.txt

If you want to save it into the same file. Drop the -i if you just want to print to stdout, which you could redirect into another file.

It will match any number of spaces followed by anything up to 2015 at the end of the line, then replace that whole match with " 2015"

Another possibility would be to do

sed -e 's/^\([[:digit:]]\{6\}\).\+\([[:digit:]]\{4\}\)$/\1 \2/'  names.txt

Which will match 6 digits at the start of the line and 4 at the end and print those matches with a space between them. It will leave any other lines unchanged.

Answer 2

my preference is to anchor to the start of the string to reduce back tracking and I like extended regexp so I would use

sed -re 's/^([0-9]{6}) .*( 2015)$/\1\2/' names.txt

what this does

-r extended regexp
-e expression to follow
s substitute
^ beginning of line
( start subpattern
[0-9] a digit
{6} previous occurs exactly six times
) end subexpression
. any single character
* previous occurs zero or more times
$ end of line
\1 first subpattern
\2 second subpattern