Clever Geek Handbook

Search for lines by pattern, leave only the pattern, but keep unmatched lines as they are

So there is a file with res.txt file like this

processing file 0x8001.values
channel 1: 123
channel 2: 234
channel 3: 345
processing file 0x8002.values
channel 1: 456
channel 2: 567
channel 3: 678

I have a template like this

0x[0-9a-f][0-9a-f][0-9a-f][0-9a-f]

Using for example

grep -o "0x[0-9a-f][0-9a-f][0-9a-f][0-9a-f]" res.txt

I can get a list of my filenames (no .values), which is good!

0x8001
0x8002

Still, all other lines that do not match the pattern remain where they are:

0x8001
channel 1: 123
channel 2: 234
channel 3: 345
0x8002
channel 1: 456
channel 2: 567
channel 3: 678

I'm pretty familiar with sed, but couldn't find a way to do what I want.

+3


source to share


2 answers


You can use sed.

$ sed 's/.*\(0x[0-9a-f][0-9a-f][0-9a-f][0-9a-f]\).*/\1/' file
0x8001
channel 1: 123
channel 2: 234
channel 3: 345
0x8002
channel 1: 456
channel 2: 567
channel 3: 678


I assumed there was only one hexadecimal string present in the string like 0x8001

.

+1


source


Using gnu awk you can use a match

function like this with your regex:



awk 'match($0, /0x[0-9a-f][0-9a-f][0-9a-f][0-9a-f]/){$0=substr($0,RSTART,RLENGTH)} 1' file
0x8001
channel 1: 123
channel 2: 234
channel 3: 345
0x8002
channel 1: 456
channel 2: 567
channel 3: 678
+1


source






More articles:


All Articles