How to delete first 10 lines containing certain string?

Question

Suppose i have a file with this structures

1001.txt
1002.txt
1003.txt
1004.txt
1005.txt
2001.txt
2002.txt
2003.txt
...

Now how can I delete first 10 numbers of line which start with '2'? There might be more than 10 lines start with '2'.

I know I can use grep '^2' file | wc -l to find number of lines which start with '2'. But how to delete the first 10 numbers of line?

You might find a clue here https://stackoverflow.com/questions/23696871/how-to-remove-only-the-first-occurrence-of-a-line-in-a-file-using-sed — JeffUK, Jul 11 '17 at 16:05

score 2 · Answer 1 · answered Jul 11 '17 at 16:07

2

You can pipe your list through this Perl one-liner:

perl -p -e '$_="" if (/^2/ and $i++ >= 10)'

answered Jul 11 '17 at 16:07

amphetamachine

27,620
12
60
72

how do I specifiy which file to check? – pwan Jul 11 '17 at 16:11

score 1 · Answer 2 · answered Jul 11 '17 at 18:02

1

Another in awk. Testing with value 2 as your data only had 3 lines of related data. Replace the latter 2 with a 10.

$ awk '/^2/ && ++c<=2 {next} 1' file
1001.txt
1002.txt
1003.txt
1004.txt
1005.txt
2003.txt
.
.
.

Explained:

$ awk '/^2/ && ++c<=2 {   # if it starts with a 2 and counter still got iterations left
    next                  # skip to the next record
} 1                       # (else) output
' file

answered Jul 11 '17 at 18:02

James Brown

36,089
7
43
59

1

Or, invert the logic: `awk '!/^2/ || ++c>10'` – William Pursell Jul 12 '17 at 00:29
i got the result just like you. but nothing change in the file. – pwan Jul 12 '17 at 12:47
It doesn't change the source file, you need to redirect the output to a new file. It's safer. – James Brown Jul 12 '17 at 12:50
how to do it? `> new_file` ? – pwan Jul 12 '17 at 13:04
Yes. `awk '...' file > new_file` – James Brown Jul 12 '17 at 13:05
Excuse me, I didn't understand what you mean? – James Brown Jul 12 '17 at 13:07
what if i want to do this more than 1 times? do I need to keep `awk '...' file > new_file` and output it to a new file? – pwan Jul 12 '17 at 13:13
than i will have n files? – pwan Jul 12 '17 at 13:14
If you want them to a single `new_file`: `awk '...' file* > new_file`. If you want them all to different files you need to script around the awk, for example: `for f in file* ; do awk '...' $f > new_$f ; done` (add appropriate quoting around the `$f` if needed). – James Brown Jul 12 '17 at 13:16
Let us [continue this discussion in chat](http://chat.stackoverflow.com/rooms/149008/discussion-between-pwan-and-james-brown). – pwan Jul 12 '17 at 13:17
@JamesBrown `awk -F '/^1/ && ++c<=2 {next} 1' '/^2/ && ++d<=1 {next} 1' test.txt > new.txt` is not working.... – pwan Jul 13 '17 at 15:49
No doubt. You have `-F` but no value for it and two different programs sequentially. Awk does not work that way. – James Brown Jul 13 '17 at 18:12
so how do I do multiple operation in only one awk. I'm trying to do the 2 operation from a file to another... – pwan Jul 15 '17 at 08:07
Untested, but I'd say: `awk '( /^1/ && ++c<=2 ) || ( /^2/ && ++d<=1 ) {next} 1'` ie. two conditions and an OR (`||`) in between. – James Brown Jul 15 '17 at 10:03

score 0 · Answer 3 · answered Jul 11 '17 at 17:08

awk alternative:

awk '{ if (substr($0,1,1)=="2") { count++ } if ( count > 10 || substr($0,1,1)!="2") { print $0 } }' filename

If the first character of the line is 2, increment a counter. Then only print the line if count is greater than 10 or the first character isn't 2.

How to delete first 10 lines containing certain string?

3 Answers3