CODEWITHSUNDEEP
regexperl
Rob John
Rob John

Reputation: 287

Regex with perl one liner

I have the following:

XXUM_7_mauve_999119_ser_11.255255

UXUM_566_mauve_999119_ser_11.255255

IXUM_23_mauve_999119_ser_11.255255

and my attempt, which did not work, at a perl one liner to extract the first digit is as follows; 

perl -pi -e "s/\S+_(\.+)_.+/Number$1/g" *.txt

I expected the following results:

Number 007
Number 566
Number 023

pls help

Upvotes: 2

Views: 790

Answers (3)

Borodin
Borodin

Reputation: 126722

The problem is that this regex pattern /\S+_(\.+)_.+/ looks for a sequence of one or more literal dots . surrounded by underscores, so something like _..._ would match, but such a sequence doesn't exist in your file. I think you didn't mean to escape the dot. But even then, because the \S+ is greedy, it would find and capture the last field delimited by underscores, and so would capture ser from all three lines. Perhaps you meant to write \d+ instead of \.+, which is pretty much what I have written below.

This will do as you ask. It looks for the first occurrence of an underscore that is followed by a number of decimal digits, and uses printf to format the number as three digits.

You can add the -i qualifier, but I suggest you test it as it is first to save overwriting your data with erroneous results. Of course you could redirect the output to another file if you wished.

perl -ne'/_(\d+)/ and printf "Number %03d\n", $1' myfile

output

Number 007
Number 566
Number 023

Upvotes: 2

piaf
piaf

Reputation: 36

cat > /tmp/test
XXUM_7_mauve_999119_ser_11.255255
UXUM_566_mauve_999119_ser_11.255255
IXUM_23_mauve_999119_ser_11.255255

perl -i -ne 'if ($_=~/^\w+\_(\d+)\_mauve/g) { printf "Number %03d\n", $1; }' /tmp/test

cat /tmp/test 
Number 007
Number 566
Number 023

Upvotes: 1

choroba
choroba

Reputation: 241758

I'd use the -n option instead of the -p option and do the printing and formatting in the code:

  perl -i~ -ne 'if (($num) = /[0-9]+/g) {
                    printf "Number %03d\n", $num;
                } else {
                    print
                }' *.txt

Upvotes: 1

Related Questions