Comparing hexadecimal values with awk - awk

Question

I'm having trouble with awk and comparing values. Here's a minimal example :
echo "0000e149 0000e152" | awk '{print($1==$2)}'
Which outputs 1 instead of 0. What am I doing wrong ? And how should I do to compare such values ?
Thanks,

Answer 1

To convert a string representing a hex number to a numerical value, you need 2 things: prefix the string with "0x" and use the strtonum() function.
To demonstrate:
echo "0000e149 0000e152" | gawk '{
print $1, $1+0
print $2, $2+0
n1 = strtonum("0x" $1)
n2 = strtonum("0x" $2)
print $1, n1
print $2, n2
}'
0000e149 0
0000e152 0
0000e149 57673
0000e152 57682
We can see that naively treating the strings as numbers, awk thinks their value is 0. This is because the digits preceding the first non-digit happen to be only zeros.
Ref: https://www.gnu.org/software/gawk/manual/html_node/String-Functions.html
Note that strtonum is a GNU awk extension

Answer 2

You need to convert $1 and $2 to strings in order to enforce alphanumeric comparison. This can be done by simply append an empty string to them:
echo "0000e149 0000e152" | awk '{print($1""==$2"")}'
Otherwise awk would perform a numeric comparison. awk will need to convert them to numeric values in this case. Converting those values to numbers in awk leads to 0 - because of the leading zero(s) they are treated as octal numbers but parsing as an octal number fails because the values containing invalid digits which aren't allowed in octal numbers, which results in 0. You can verify that using the following command:
echo "0000e149 0000e152" | awk '{print $1+0; print $2+0)}'
0
0

Answer 3

When using non-decimal data you just need to tell gawk that's what you're doing and specify what base you're using in each number:
$ echo "0xe152 0x0000e152" | awk --non-decimal-data '{print($1==$2)}'
1
$ echo "0xE152 0x0000e152" | awk --non-decimal-data '{print($1==$2)}'
1
$ echo "0xe149 0x0000e152" | awk --non-decimal-data '{print($1==$2)}'
0
See http://www.gnu.org/software/gawk/manual/gawk.html#Nondecimal-Data

Answer 4

i think many forgot the fact that the hexdigits 0-9 A-F a-f rank order in ASCII - instead of wasting time performing the conversion, or risk facing numeric precision shortfall, simply :
trim out leading edge zeros, including the optional 0x / 0X
depending on the input source, also trim out delimiters such as ":" (e.g. IPv6, MAC address), "-" (e.g. UUID), "_" (e.g. "0xffff_ffff_ffff_ffff"), "%" (e.g. URL-encoding) etc
—- be mindful of the need to pad in missing leading zeros for formats that are very flexible with delimiters, such as IPv6
compare their respective string length()s :
if those differ, then one is already distinctly larger,
— otherwise
prefix both with something meaningless like "\1" to guarantee a string-compare operation without risk of either awk being too smart or running into extreme edge cases like locale-specific peculiarities to its collating order :
(("\1") toupper(hex_str_1)) == (("\1") toupper(hex_str_2))