有什么办法可以消除sed第一次出现的模式？

Question

This is a test file:

xxx
hello 
hello 
others

I want to use sed:

sed -i '/hello/d' the_test_file

This will remove all lines contain "hello", so how to just remove the first line contains "hello"?

I think sed cant do it, but with perl i can. it's like:

perl -00 -pe "s/hello//" the_test_file > /tmp/file && mv /tmp/file the_test_file

Answer 1

Show the original content of the file.

cat the_test_file
1
2
3
popo
hello 1
hello 2
others

Show the result after running the sed command

sed -e '0,/^hello/{//d;}' the_test_file
1
2
3
popo
hello 2
others

To finally answer your question

NOTE: thanks to jaypal

sed -i"bak" -e '0,/^hello/{//d;}' the_test_file

Answer 2

To remove the first hello

sed -e "/hello/ N;n;d;" the_test_file

To remove the second hello

sed -e "/hello/ N;n;n;d;" the_test_file

If there are more than 2 lines of hello consecutively, and want to retain the first hello and delete the rest

sed -e ":loop" -e "/hello/ N; s/\(hello.*\)\n\(hello.*\)/\1/" -e "t loop" the_test_file

If there are more than 2 lines of hello consecutively, and want to retain the last hello only

sed -e ":loop" -e "/hello/ N; s/\(hello.*\)\n\(hello.*\)/\2/" -e "t loop" the_test_file

the_test_file

xxx
hello1
hello2
hello3
others

Answer 3

awk solution:

$ cat file
something
hello 1
else
hello 2
others
hello 3

$ awk '/hello/ && ++c==1{next}1' file
something
else
hello 2
others
hello 3

Answer 4

This does the trick (but doesn't generalize easily):

$ sed -e '1,/hello/p' -e '1,/hello/d' -e '/hello/d' <<!
> xxx
> hello 
> hello 
> others
!
xxx
hello 
others
$

The logic is to print the lines from line 1 to the first line containing 'hello'; then delete those lines (so the automatic print does not happen); thereafter, simply delete the lines matching 'hello'.