Awk (gnu awk) printing comparisons with and without parens have different results

Question

awk 'BEGIN { print 3 > 2 }' a.out

Prints nothing

But

awk 'BEGIN { print (3 > 2) }' a.out
> 1

Prints 1. This repros with any file. Why do the parenthesis have an effect?

GNU AWK 5.1

Answer 1

awk has Redirection like the shell. Check your current directory for a file named "2" -- it will have one line containing the value "3".

With parentheses, you're changing how awk parses the code: giving the expression 3 > 2 higher priority, you're forcing awk to use the relational operator.

This is discussed in the Precedence chapter of the manual:

< <= == != > >= >> | |&

Relational and redirection. The relational operators and the redirections have the same precedence level. Characters such as '>' serve both as relationals and as redirections; the context distinguishes between the two meanings.

See the index of the gawk manual for all uses of >

Answer 2

that's why, whenever i need x greater-than y compares in awk , the safer approach is reverse it and write

y < x

that way, only getline ones need to be watched out for, as opposed to accidentally overwriting files

'{ print x > y }'

that happen to have the same name as y 's string-value. side note - to best of my understanding, the cleanest way to force numeric compare is probably

+y < +x       # doing ( 0+y < 0+x ) is superfluous 
              # in this context since unary `+` has an 
              # implicit `0` to left of it