Shell Programming and Scripting

This might work on some systems, but it certainly is not portable.

The standards state that there is no associativity for the == operator and some versions of awk produce the syntax error:

Code:

awk: syntax error at source line 1
 context is
	 >>> $3==$4== <<< 
awk: bailing out at source line 1

If we rewrite the expression as:

Code:

awk -F, '$3==($4==($5==1))'

then there are lots of cases where that expression will evaluate to 1 even if all three of those fields are not set to 1. For example, the above command will print any of the following lines:

Code:

a,b,1,1,1
a,b,1,0,X for any X other than 1
a,b,0,1,X for any X other than 1
a,b,0,W,X for any W other than 0 or 1 for any X

Of course, it could also be rewritten as:

Code:

awk -F, '(($3==$4)==$5)==1))'

which would print any of the following lines:

Code:

a,b,1,1,1
a,b,X,X,1 for any X
a,b,X,Y,0 for any X that is not Y

Note that grep will work as well as egrep (or the preferred syntax grep -E) for the RE being used in this thread.

Note also that the RE suggested works correctly only if there are exactly 6 fields (separated by 5 commas) on each input line. Since BREs and EREs use a greedy match, the RE .*, can match more than one field if there are more than 5 commas on a line. For example, that egrep command will also print the lines:

Code:

a,b,c,1,1,1,2
a,b,0,0,1,1,1,2
1,2,1,2,1,2,1,2,1,2,1,1,1,2

in addition to lines with 1 in fields 3,4, and 5 that only have 6 fields.

To make it work correctly on a line containing six commas (i.e. 7 fields), you would need to change the RE to:

Code:

.*,.*,1,1,1,.*,

and you would need to add an additional .*, to the end of that RE for each additional field in your input file.

Alternatively, we could use an RE that only matches non-comma characters in each of the first two fields:

Code:

grep '^[^,]*,[^,]*,1,1,1,' input_file

which will only print lines with 1 in fields 3, 4, and 5 as long as there are at least six fields on each line. ([^,]* is an RE that matches zero or more occurrences (*) of any character that is not a comma ([^,]) followed by a comma (,). And, the leading ^ in the entire RE anchors the match to the start of the line.)