Grepping something out of a specific column in a file using pattern from another file

[BBAF]

Hello,
I often use grep to extract lines from a file using patterns that are in another file, e.g.

grep -wf patternfile.txt infile.txt

This toy example illustrates my problem:

Infile:
A 100 3396 101 M
A 200 3488 100 M
B 100 3431 102 M
A 100 3454 121 F
C 200 3407 378 M
A 200 3440 400 M


Patternfile:
100
121
378
407

Here, I want grep to only look for matches in column 4 of the infile. So, the output that I want is:

A 200 3488 100 M
A 100 3454 121 F
C 200 3407 378 M

However, since grep also finds some matches in column 2 of the infile, what I get is:
A 100 3396 101 M
A 200 3488 100 M
B 100 3431 102 M
A 100 3454 121 F
C 200 3407 378 M

In this case I can get around it by, e.g.
awk '{print $3,$4,$5}' infile.txt | grep -f patternfile.txt > temp.txt
grep -f temp.txt infile.txt

However, I would much more like to be able to grep directly out of a specific column with a pattern from a file, as this workaround isn't always applicable.

I'd appreciate any suggestions

[radoulov]

A few possibilities:

1. Modify your pattern file, use the following format:

([^ ]+ ){3}100
([^ ]+ ){3}121
([^ ]+ ){3}378
([^ ]+ ){3}407

Then use:

grep -Ef patternfile infile

2. Construct the pattern with a process subsitution (if your shell supports it):

&#37; grep -Ef <(printf "([^ ]+ ){3}%s\n" $(<patternfile)) infile
A 200 3488 100 M
A 100 3454 121 F
C 200 3407 378 M

3. Use a more powerful tool:

% awk 'NR==FNR{_[$1];next}$4 in _' patternfile infile 
A 200 3488 100 M
A 100 3454 121 F
C 200 3407 378 M

[BBAF]

Thanks radoulov! Just what I needed. I had a hunch that awk could do the job - and I think that's the best of your solutions.