Grep a file pattern in another

Shell Programming and Scripting
1

Hi

I'm new to the forum, so I'd apologize for any error in the format of the post.

I'm trying to find a file content in another one using:

grep -w -f file1 file2

file1

GJA7
TSC

file 2

GJC1	GJA7
TSC1	TSC
TSC22D3	TSC-22R

so I expect the result to be:

GJC1	GJA7
TSC1	TSC

but instead it's:

GJC1	GJA7
TSC1	TSC
TSC22D3	TSC-22R

actually the problem is that it sees TSC-22R like two words, not one, and so matches the first part it with TSC.

Thanks

2

Try

$ cat a
GJA7
TSC

$ cat b
GJC1    GJA7
TSC1    TSC
TSC22D3    TSC-22R

$ awk 'NR == FNR {_[$0]++; next} !($0 in _) && length($1)<5' a b
GJC1    GJA7
TSC1    TSC
3

Try:

awk 'NR==FNR{A[$1]; next} {for(i=1; i<=NF; i++) if ($i in A) {print; next}}' file1 file2
1 Like
4

you missed 2nd file I think while posting:)

I think second if ($i in A) {print; next}} is not necessary

Scrutinizer do you have solution for this thread using grep I am interested

1 Like
5

That is right, thanks I corrected it in my post, thanks. The {print; next} is there, otherwise a line would get printed twice if there is a double match, for example..

6

I was not aware about it...Thank you..so much Scrutinizer

7

A grep solution might be (if your grep understands "-"):

sed 's/^/(^|[[:space:]])/; s/$/([[:space:]]|$)/' file1 | grep -Ef - file2

or (in bash / ksh93 )

grep -Ef <(sed 's/^/(^|[[:space:]])/; s/$/([[:space:]]|$)/' file1 ) file2
1 Like
8

It worked fine...Thanks a lot Scrutinizer

9

Hi,
Another way with grep and xargs (as sed Scrutinizer solution):
1) xargs posix:

grep -Ef <(cat file1 | xargs -I {} printf "(^|[[:space:]]){}([[:space:]]|$)\n") file2

2) xargs gnu:

grep -Ef <(xargs -a file1 -I {} printf "(^|[[:space:]]){}([[:space:]]|$)\n") file2

Regards.

1 Like
10

Every grep suggestion thus far in this thread (including the OP's) assumes that it is safe to evaluate file1 in a regular expression context. If file1 does not contain regular expressions (very likely), and if its literal text may contain regular expression metacharacters (possible), that assumption would be unjustified and could yield erroneous output.

The assumption may be valid (probably is), but given the miniscule data sample and lack of specifics in the OP, I thought it prudent to mention it.

Regards,
Alister

1 Like
11

I agree, and with these grep suggestions there is then also sensitivity to spacing in file1 (one leading or trailing space somewhere in the file will influence the result, which may not be desirable) so generally I would prefer the awk approach.

12

Thank you guys! You're just fantastic! It resolved my problem.
Best