grep many different strings

Question

I have a file strings.txt containing 100 strings, each on a line

string1
string2
...
string100

For each of those strings, I want to find all the lines in file_to_look.txt that contain that string. Now, I could run grep 100 times like grep string1 file_to_look.txt, then grep string2 file_to_look.txt, etc., but that would take a lot of typing time for me.

Is there a way that I don't have to do so much typing?

EDIT: Solutions that go through file_to_look.txt only 1 time instead of 100 times would be great, since my file_to_look.txt is quite large.

score 4 · Accepted Answer · answered Nov 06 '12 at 22:27

4

-f is for passing (GNU) grep a pattern file.

grep -f strings.txt file_to_look.txt

answered Nov 06 '12 at 22:27

j.w.r

4,136
2
27
29

This seems like the most elegant way to do it; +1 – sampson-chen Nov 06 '12 at 22:39
Does it go through the input file 1 time or 100 times? – Mika H. Nov 06 '12 at 22:41
It reads the pattern file once, then scans the input file once, so just 1 time. – j.w.r Nov 06 '12 at 22:48

Ian Kelling · Answer 2 · 2012-11-06T22:26:41.680

0

while read line; do grep "$line" file_to_look.txt; done < strings.txt

That does exactly what you asked for. Substitute ";" for newlines as you see fit.

xargs is the other option people will suggest. My suggestion is to generally look for another alternative first, as xarg has a bunch of gotchas which can make things go very wrong.

Greg's bash wiki on xargs

edited Nov 06 '12 at 22:26

answered Nov 06 '12 at 22:20

Ian Kelling

9,643
9
35
39

score 0 · Answer 3 · answered Nov 06 '12 at 22:22

0

Usually xargs is used for repeating commands with multiple values:

xargs -I{} grep {} file_to_look.txt < strings.txt

answered Nov 06 '12 at 22:22

BeniBela

16,412
4
45
52

score 0 · Answer 4 · edited May 23 '17 at 10:24

0

You can do it using while read, like this:

cat strings.txt | while read line ; do grep "$line" file_to_look.txt ; done

For more alternative ways take a look at:

How to apply shell command to each line of a command output?

edited May 23 '17 at 10:24

Community

1
1

answered Nov 06 '12 at 22:24

Nelson

49,283
8
68
81

score 0 · Answer 5 · answered Nov 06 '12 at 22:32

You can do it with the following short script:

#!/bin/bash

file_name=string.txt
file_to_look=file_to_look.txt
patterns=$(tr '\n' '|' $filename)
patterns=${patterns%?} # remove the last |
grep -E $patterns $file_to_look

This rounds up all of your search patterns together and hands it off to grep in one go with the -E option, so grep only has to parse through file_to_look.txt once, instead of a 100 times.

grep many different strings

5 Answers5