filtering rows using awk

Question

I have a file like this which is tab separated:

name    v1  v2  v3  v4
g1  4.5 2.3 2.1 0.2
g2  10  3   5   2.3
g3  7   2.5 2.8 3.9

Just showing you above a dummy file where I have 5 columns and 4 rows(including the header). I want to filter out the rows such that if every column in a particular row has value >= 2, then keep that row else remove it. The output should look like this:

name    v1  v2  v3  v4
g2  10  3   5   2.3
g3  7   2.5 2.8 3.9

How can I do it using awk?

Answer 1

AFAIK awk doesn't have a way of doing this short of iterating over fields explicitly. For example:

$ awk 'NR>1 {for(i=2;i<=NF;i++) if($i+0 < 2) next} 1' file
name    v1  v2  v3  v4
g2  10  3   5   2.3
g3  7   2.5 2.8 3.9

Answer 2

steeldriver has already provided an awk solution. Here's a perl version (using an array slice rather than a for loop):

$ perl -lane 'print if ($.==1 || grep ($_ >= 2, @F) == $#F)' input.txt
name    v1  v2  v3  v4
g2  10  3   5   2.3
g3  7   2.5 2.8 3.9

This prints only the first (header) line and lines where all numeric fields have a value greater than or equal to 2. (non-numeric fields like g1 or g2 will evaluate to 0)

---

Note: perl's grep() function is similar in concept, but not exactly the same as the grep command-line program.

grep(expression,array) runs the expression in its first argument (e.g. $_ >= 2) against every element of an array (e.g. @F), and returns an array consisting of every element where the result was true.

In a scalar context (such as a numeric comparison against an integer), it returns the number of times the expression was true, instead of an array. This is what we are doing here with == $#F, to test equivalence with $#F (the number of elements in array @F)

The expression can be a simple test as used in this example, or a code block containing any perl code. It can also optionally modify each element. e.g. @new = grep(s/foo/bar/g, @old) would populate @new with all elements from @old which were successfully modified (i.e. thost that contained at least one "foo". all of which were changed to "bar"). See perldoc -f grep for details.