GNU sed 4.2.1 matching second occurence

Question

With this command I am trying to insert M09 between two patterns of M502 and the second line since M502 that looks like X41.5Y251.5T201.

Specifically I am trying to insert M09 before the last line.

sed -i "/M502/{:a;N;/\(.*\)T\([0-9]\+\)/!ba;N;s/\(.*\)\n\(.*\)T\([0-9]\+\)/\1\nM09\n\2T\3/}" *.nc

Result:

M502
M09
X1287.2Y353.28T324
..........
G27X-310.27
X41.5Y251.5T201

Should be:

M502
X1287.2Y353.28T324
..........
G27X-310.27
M09
X41.5Y251.5T201

I am trying to match the last line with \(.*\)T\([0-9]\+\), but that picks up the second line. How do I make sed ignore the first match of this expression?

I stress that I could have an undertermined amount of such text blocks and I need to insert M09 for every one.

RavinderSingh13 · Answer 1 · 2021-10-26T12:25:31.883

With your shown samples, please try following awk code.

awk '
/.*T[0-9][0-9]*/{ found=1       }
FNR==1          { prev=$0; next }
                {
                  print prev
                  prev=$0
                }
END{
  if(found)     { print "M09"   }
  print prev
}
' Input_file

Above code will only print the values on terminal, in case you want to save output into Input_file itself then append > temp && mv temp Input_file to above code.

Andreas Louv · Answer 2 · 2021-10-26T12:28:06.323

1

How do I match the second instance of \(.*\)T\([0-9]\+\) ? It must be an expression. Thanks for the help.

Assuming that you would like to insert some text just before the line that matches the the above regex for the 2th time, then you should consider using awk, as then you could do something like this:

$ awk '/.*T[0-9][0-9]*/ { n++ } !p && n == 2 { print "M09"; p = 1 } 1' input.txt
M502
X1287.2Y353.28T324
..........
G27X-310.27
M09
X41.5Y251.5T201

The breakdown of the above command is:

/.*T[0-9][0-9]*/ { n++ }            # Everytime the regex match add one to `n`
!p && n == 2 { print "M09"; p = 1 } # If n = 2 and we havn't printed a header yet,
                                    # then print "M09"
1                                   # Print every line

edited Oct 26 '21 at 12:28

answered Oct 26 '21 at 12:20

Andreas Louv

46,145
13
104
123

As I understand this would add M09 once before the second "global" instance of .*T[0-9][0-9]*. I can have an undertermined number of text blocks where I need to add M09. So with my sed example I just need to ignore the first \(.*\)T\([0-9]\+\) match after every M502. – Vardenis Pavardenis Oct 26 '21 at 13:53
@VardenisPavardenis You need to edit your question, or post another one entirely with your specific problem. As it's evident from the amount of different answers that your question isn't clear. – Andreas Louv Oct 26 '21 at 14:33
It is all my specific problem. Edited it to try and express it more clearly. – Vardenis Pavardenis Oct 27 '21 at 05:32

GNU sed 4.2.1 matching second occurence

2 Answers2