每行运行命令？

Question

I have a bunch of 1255-138279.trans.txt files that contain lines like these:

1255-138279-0006 THE SHAPE WENT SLOWLY ALONG BUT WITHOUT MUCH EXERTION FOR THE SNOW THOUGH SUDDEN WAS NOT AS YET MORE THAN TWO INCHES DEEP
1255-138279-0007 AT THIS TIME SOME WORDS WERE SPOKEN ALOUD ONE
1255-138279-0008 TWO THREE
1255-138279-0009 FOUR FIVE
1255-138279-0010 BETWEEN EACH UTTERANCE THE LITTLE SHAPE ADVANCED ABOUT HALF A DOZEN YARDS
1255-138279-0011 HERE THE SPOT STOPPED AND DWINDLED SMALLER
1255-138279-0012 THEN A MORSEL OF SNOW FLEW ACROSS THE RIVER TOWARDS THE FIFTH WINDOW
1255-138279-0013 THE RIVER WOULD HAVE BEEN SEEN BY DAY TO BE OF THAT DEEP SMOOTH SORT WHICH RACES MIDDLE AND SIDES WITH THE SAME GLIDING PRECISION ANY IRREGULARITIES OF SPEED BEING IMMEDIATELY CORRECTED BY A SMALL WHIRLPOOL
1255-138279-0014 THE WINDOW WAS STRUCK AGAIN IN THE SAME MANNER
1255-138279-0015 THEN A NOISE WAS HEARD APPARENTLY PRODUCED BY THE OPENING OF THE WINDOW
1255-138279-0016 SAID THE BLURRED SPOT IN THE SNOW TREMULOUSLY
1255-138279-0017 I ASKED WHICH WAS YOUR WINDOW FORGIVE ME
1255-138279-0018 WELL I SAID THAT YOU MIGHT
1255-138279-0019 O MUST I IT IS WHEN SHALL WE BE MARRIED FRANK
1255-138279-0020 I HAVE MONEY
1255-138279-0021 AND WE LIVE IN TWO PARISHES DO WE WHAT THEN
1255-138279-0022 IF I SAID SO OF COURSE I WILL
1255-138279-0023 THE FACT IS I FORGOT TO ASK
1255-138279-0024 GOOD NIGHT FRANK GOOD NIGHT

For each line, I want to create a new file, named 1255-138279-0008.txt (or whatever the code at the start of the line is), and containing the remainder of that line (in this case TWO THREE ).

Answer 1

This simple while loop in shell would do the job:

while read -r id line; do
   echo "$line" >> "$id.txt"
done < file

Answer 2

Try this with awk :

$ awk '{v=$1; $1=""; print $0 > "path/to/dir/"v".txt"}' file.txt
$ ls -1 path/to/dir/*.txt

Answer 3

I prefer doing this in two steps. First, use sed or awk to generate a shell script:

sed "s/^\s*\(\S\+\)\s\+\(.*\)$/echo '\2' > '\1.txt'/g" input.txt > output.sh

Or with a POSIX-compliant regex (without GNU extensions):

sed "s/^[[:space:]]*\([^[:space:]]\+\)[[:space:]]\+\(.*\)$/echo '\2' > '\1.txt'/g" input.txt > output.sh

Or using ERE:

sed -E "s/^[[:space:]]*([^[:space:]]+)[[:space:]]+(.*)$/echo '\2' > '\1.txt'/g" input.txt > output.sh

Output:

...
echo 'TWO THREE' > '1255-138279-0008.txt'
echo 'FOUR FIVE' > '1255-138279-0009.txt'
...

Then execute that script to generate the files.

. output.sh