How do I cut letter X out of a word?

问题

I need to cut letter X out of a word: For example: I need to cut the first letter out of Star Wars, the fourth out of munich,...

1 star wars
4 munich
5 casino royale
7 the fast and the furious
52 a fish called wanda

tar wars
munch
casio royale
the fat and the furious
a fish called wanda

I already tried it with cut, but it didn't work. This was my command: sed 's/^\([0-9]*\) \(.*\)/ echo \2 | cut -c \1/'

So it gave me this output:

 echo star wars | cut -c 5
 echo munich | cut -c 5
 echo casino royale | cut -c 5
 echo the fast and the furious | cut -c 5
 echo a fish called wanda | cut -c 52

And than if I send it to bash. I only get the X th letter of the word.

I need to do the exercise with sed and other commands. But I can't use awk or perl.

Thanks

回答1:

This might work for you (GNU sed):

sed 's/^\([0-9]*\) \(.*\)/echo '\''\2'\''|sed '\''s\/.\/\/\1'\''/e' file

This uses the e flag of the s command to evaluate the RHS and runs a second sed invocation using the backreferences from the LHS. Perhaps easier on the eye is this:

sed -r 's/^([0-9]*) (.*)/echo "\2"|sed "s#.##\1"/e' file

回答2:

You can use just bash and its parameter expansion:

while read n s ; do
    echo "${s:0:n-1}${s:n}"
done < input.txt

If you need one line, just remove the newlines and add a semicolon:

while read n s ; do echo "${s:0:n-1}${s:n}" ; done < input.txt

If you really need to use sed and cut, it's also doable, but a bit less readable:

cat -n input.txt \
| sed 's/\t\([0-9]\+\).*/s=\\(.\\{\1\\}\\).=\\1=/' \
| sed -f- <(sed 's/[0-9]*//' input.txt) \
| cut -c2-

Explanation:

number the lines
turn each line into a sed command that searches for the given number of characters and removes the one following them
run the generated sed command on the original file with the numbers removed
remove the extra leading space

回答3:

you can use sed in this way:

sed -e '1s/\([a-z]\)\{1\}//' -e 's/^[0-9]\+\s\+\(.*\)/\1/g' file.txt

The first sed regular expresion works on the first line and replace the first character and the second regular expresion works with rest of text, 1 column: one number or more, 2 column; one space or more, after this I put the remaining test in one match \(.*\) and replaced all with this match.

来源：https://stackoverflow.com/questions/34591564/how-do-i-cut-letter-x-out-of-a-word

标签

bash

sed

cut