CODEWITHSUNDEEP
unixsedheader
Carla
Carla

Reputation: 11

using sed to remove part of a header line starting with forward slash

I am trying to adjust header lines in a file using sed and I have looked at many previous posts but nothing I've tried based on those so far have helped. So here it goes: I am using terminal on mac OS, I have a file with header lines that look like this: starting with >

>KeepThis_text_VariableNumbersHere /LotsOFText/Here to get rid of."

I want to have:

>KeepThis_text_VariableNumbersHere

All of the header lines end in ." but there are quotes within the header line as well.

I have tried many variations using sed, here are a few:

 sed -e 's/^\/*.*//' input.file > output.file  #this removed everything

or

 sed -e 's/^\/*.*"//' input.file > output.file #this kept non-header

lines, which is good, but got rid of entire header line

sed -e 's/>KeepThis_Text_*.* *.*”/>KeepThis_Text_*.*\//' input.file > output.file #This did not change anything

Thank you for any suggestions!

Upvotes: 0

Views: 1171

Answers (4)

karakfa
karakfa

Reputation: 67497

awk to the rescue!

awk -F/ 'NR==1{print $1;next}1' file

modifies first line only, prints everything up to /.

Upvotes: 0

hek2mgl
hek2mgl

Reputation: 158060

I would use capture everything starting from > including everything before space followed by a / in a capturing group. The keep only the capturing group and throw away the remaining text, like this:

sed 's/\(>[^\/]\+\) .*/\1/' input.txt

It looks more clean using extended regular expression, using the -r option:

sed -r 's/(>[^\/]+) .*/\1/' input.txt

I'm not sure if header line means for you the first line. If you really mean the first line, limit the s command to the first line like this:

sed -r '1s/(>[^/]+) .*/\1/' input.txt

Upvotes: 0

F. Hauri - Give Up GitHub
F. Hauri - Give Up GitHub

Reputation: 70852

If you only want to change 1st line, consider addressing your command (s///):

sed -re '1s/\/\w+//g'

or addressing all lines beginning with >:

sed -re '/^>/s/\/\w+//g'

or if you want to drop everything from / to end of line:

sed -re '/^>/s/\/.*//g'

Upvotes: 1

jxh
jxh

Reputation: 70452

If you want to only keep the part that precedes the forward slash, then the following will do:

sed -e 's-/.*--'

The separators for the command are not limited to /, it is just a convention. Here, I used - as command separators. The command is then: replace anything following a slash until the end of line with empty text.

Upvotes: 1

Related Questions