sed and awk – How to Extract Line from File on Specific Condition

awksed

File:

chromosome  position  ref  alt 
chr1          1398     A    T 
chr1          2980     A    C 
chr2          3323     C    T,A
chr2          3749     T    G
chr3          5251     C    T,G
chr3          9990     G    C,T
chr4          10345    T    G

I need to extract full line when column 4 has 2 or more characters separated by comma

Expected Output is:

chr2          3323     C    T,A
chr3          5251     C    T,G
chr3          9990     G    C,T

Best Answer

A couple of other ways to look at this.

Method #1

Since you only are interested in lines if they have more than 2 characters separated by commas you could just grep for commas:

$ grep "," sample.txt 
chr2          3323     C    T,A
chr3          5251     C    T,G
chr3          9990     G    C,T

Method #2

You could use grep's PCRE facility. This is where grep can use Perl's regular expression engine to do the matching. It's quite powerful and lets you do a lot of what you can do with Perl from grep.

loosely defined

$ grep -P "(\w,)+" sample.txt

strictly defined

$ grep -P '\w+\d\s+\d+\s+\w\s+(\w,)+' sample.txt

Method #3

Using awk. This again is taking advantage of the fact that only the lines with a comma (,) are of interest, so it just finds them and prints them:

loosely defined

$ awk '/,/{print}' sample.txt

more strictly defined

$ awk '/([[:alpha:]])+,[[:alpha:]]/{print}' sample.txt

even more strictly defined

$ awk '$4 ~ /([[:alpha:]])+,[[:alpha:]]/{print}' sample.txt

This one looks at the contents of the 4th column and checks that it's a letter followed by a comma, followed by another letter.

even more strictly defined

$ awk '$4 ~ /([GATC])+,[GATC]/{print}' sample.txt

This looks for only a G,A,T, or C followed by a comma, followed by another G,A,T or C.

Related Question