UNIX for Dummies Questions & Answers

On Solaris 10:

Code:

 nawk -v f=f 'BEGIN{printf("%u %u %u %u %u\n", "0x"f, 0+("0x" f),"0x" f) + 0, "0x"f+0, 0+"0x"f);exit}'
0 0 0 0 0

So nawk does do what is wanted.

No, what is wanted is 15. We want 0+"0xf" to be treated as a number with hexadecimal value 0xf that is to be printed as the unsigned decimal number 15.

Thanks - my bad.

This, however does produce correct output.

Code:

$ printf "%d\n" 0x0362de20  "0x"0362de20
56811040
56811040

1. this emulates your approach - prepending "0x" to a hex number
2. it does produce correct output, I believe.
3. So, is it feasible to consider printf
sed or awk to pipe (or process substitution) prepended data to printf ?
data to printf?
Is there some reason we must only use awk to output

The submitter's original script used cat once and awk twice before doing the conversion from hex to decimal. I cut that back to just one awk, but clearly there is no need for awk here either. We can do it all with shell built-ins...

Code:

/usr/xpg4/bin/sh
while read f1 f2 junk f4 f5 f6 f7 f8 f9 f10 junk
do	[ "$f1" = "" ] && continue
	printf '%s %s %u %u %u %u %u %u %u\n' \
	    "$f1" "$f2" "0x$f4" "0x$f5" "0x$f6" "0x$f7" "0x$f8" "0x$f9" "0x$f10"
done < hexconv

In theory, this script should work with /bin/sh, /bin/ksh, /bin/bash, and /usr/xpg4/bin/sh, but I don't know if the printf built-in in all of those shells will recognize hexadecimal string values (and Jim didn't say which shell he was using for the results he showed us in post #10 in this thread).

Hello Don, Finally I could get the results now

Let's compare your code (reformatted to be more easily seen on a normal screen):

Code:

cat hexconv |
awk 'NF > 0' |
while read a b c d e f g h i j k l m n o
do     printf "%s %s %d %d %d %d %d %d %d\n" \
	    "$a" "$b" "0x$d" "0x$e" "0x$f" "0x$g" "0x$h" "0x$i" "0x$j"
done

and the code I suggested:

Code:

while read f1 f2 junk f4 f5 f6 f7 f8 f9 f10 junk
do	[ "$f1" = "" ] && continue
	printf '%s %s %u %u %u %u %u %u %u\n' \
	    "$f1" "$f2" "0x$f4" "0x$f5" "0x$f6" "0x$f7" "0x$f8" "0x$f9" "0x$f10"
done < hexconv

Your code requires that a shell be exec'ed to run your script, that cat be exec'ed to read your input file, and that awk be exec'ed to discard empty lines found in your input file.

My code requires that a shell be exec'ed to run my script. Note that the exec operation is a relatively EXTREMELY slow operation on UNIX and Linux systems.

Your script has cat read the entire contents of your input file, write the entire contents of your input file, has awk read the entire contents of your input file and write all non-blank lines in your input file, and has the shell read all non-blank line in your input file and write parts of those non-empty lines in a different format.

My script reads your input file once and writes parts of non-empty lines in a different format.

Your script reads 15 fields (named a through o) and never uses $c, $k, $l, $m, $n, or $o.

My script reads the same data, but uses the name junk for data that will not be used by the script. It uses ffield# as the names of the fields it does use. (If headings had been supplied on the sample input, I would have used variables names related to the field names.)

Your script uses %d as the format specifier when printing the converted hexadecimal values. My uses %u. By definition, hexadecimal values are unsigned. But for the sample data provided, the output will be identical. If some of your input data contains a sixteen hexadecimal digit value with the highest order hexadecimal digit being 8 or greater, your script will print it as a negative decimal value while my script will print it as a positive decimal value.

You seem to believe that:

Code:

cat hexconv |  awk 'NF > 0' |

is needed to get rid of the empty lines in your input file. I believe that:

Code:

[ "$f1" = "" ] && continue

is a much simpler and much more efficient way to ignore empty lines in your input file and it still would be if you used the much more efficient:

Code:

awk 'NF > 0' hexconv |

without the cat.

Unfortunately, now that I've tried to answer your original question:
about the complexity of your code, you have changed the question to:
and I don't know what that means. How do you subtract MET with a hexadecimal value like 00000000fbfa0000? And, if you did subtract field 3 from field 4 or subtract field 4 from field 3, what different output are you hoping to produce as a result? Does the difference replace field 3 or field 4, or is another output field supposed to be created?