CODEWITHSUNDEEP
sedword-count
Amir
Amir

Reputation: 25

How to insert total number of lines in a new line before the first line command?

I have thousands of .xyz files which are chemical coordinates like this one (for instance):

  Fe        0.000000000     0.000000000     0.000000000
   C        2.112450000     0.000000000     0.000000000
   C        0.039817193     1.817419422     0.000000000

I searched a lot for a simple command, like sed or head and tail, to write the counted number of lines on top of the file (with making a newline \n and with two spaces before the total number) but couldn't be successful. I would really appreciate any help given. The output mst be like this:

  3
Fe       0.000000000     0.000000000     0.000000000
C        2.112450000     0.000000000     0.000000000
C        0.039817193     1.817419422     0.000000000

Upvotes: 0

Views: 140

Answers (2)

SLePort
SLePort

Reputation: 15461

Try this:

sed '1i \
'$(wc -l < file.xyz) file.xyz

To process multiple files with find and edit the files in place using the -i flag:

find . -name '*.xyz' -exec sh -c 'sed -i "1s/.*/$(wc -l < {})\n&/" {}' \;

NB: as the files will be edited in place with this command, you may want to first check the output. To do so, remove the -i flag after sed. If the command meets your needs, then add the -i flag again.

Upvotes: 1

John Zwinck
John Zwinck

Reputation: 249153

You can do it as a composite command:

(wc -l < fileA && cat fileA) > outputA

Note that I used < there to make sure wc does not print the filename on the first line. On Mac OS at least, it does if you don't use redirection like that.

Edit: If you need to apply the same to many files:

mkdir output
ls *.xyz | while read filename; do
    (wc -l < $filename && cat $filename) > output/$filename
done

Just for fun, here's a command that you should not use, but works in some cases:

tee xxx < fileA | wc -l > xxx # don't do this

Upvotes: 1

Related Questions