I've written the following script to extract a certain numeric field after reading a file (e.g. 50 lines long) line by line, and then perform a simple subtraction on that particular field using bc. This particular line does not seem to work and I have a hard time understanding why. Any help much appreciated!

PS. I could use awk to do it but would like to know what's wrong with bc in this case

#! /bin/sh

start=1
step=1
end=50
#input=datafile with 50 lines

for data in `seq $start $step $stop`
do

line=`cat datafile | gawk -vcount="$data" 'NR=count {print}'`
# this prints the line that's being read

r=`echo $line | gawk '{print $2}'`
# extracts the field, I can echo r and get the desired output

radius=`echo "300 - $r" | bc`
# this line doesn't work, why?
done

A couple of small problems...

In the script, you have declared end=50, and then below used "seq $start $step $stop" - so it looks like you meant to use $end instead of $stop, which was so far undefined.

And then in the first gawk, I think that 'NR=count {print}' works better as 'NR==count {print}' - that extra "=" sign is necessary for an equality.

And because that first gawk statement isn't working properly, the values of $line and $r are not as expected either, leading to an error from bc. At least, that's how it appears to me.

Such that finally, when that gawk was fixed, then the script basically works.Happy with ur solution... then tick "yes" and mark as Solved!

Please use ***[code][/code]*** tags around your code and data, to preserve the original formatting and to improve readability. Do not use quote tags, bolding, colors, "start/end" lines, or other creative techniques.


Does this have to be a posix-compatible sh script, or can you use bash or another more advanced shell with features that can make things easier?


In any case, the first bit of advice I have is that $(..) is highly recommended over `..`. This is a fully posix-compatible feature, so there's really no reason to ever use backticks anymore.

Second, formatting! Clean, consistent formatting makes code readable and more easily debuggable. Be liberal with whitespace; indent all your sub-commands evenly, and separate logical sections with empty lines. Never just line everything up on the left edge. Add comments anywhere the code isn't completely obvious (and remember, what seems obvious to you now will not be so a year or so down the line).

One non-posix feature I would recommend is to use the c-style for loop, instead of seq.
If you simply want to read every line of the file, use a while+read loop instead.

Now a few more notes about the original script:

1)
Another option, instead of a for loop, would be a while loop:

2)
It's not a good idea to hard-code filenames and similar data sources in a script.
Set it up in a variable at the top and use the variable in the code itself.


3)
as mentioned, in awk "=" sets a variable, "==" tests it.
Also notice the Useless Use Of Cat.

sed is easier to use when extracting lines anyway.
4)
If using bash, a herestring can be used instead of echo.
And cut can usually be used instead of awk for simple field extraction.

Built-in parameter substitution may be even better.

(This example is a guess since I don't know what the line actually looks like,
but assuming space-delimited fields...)

There are other string manipulation techniques available too, it the above isn't satisfactory. Using read, for example:


But as I showed above, the whole of the above section can be condensed into a single awk command anyway.

1 members found this post helpful.

Dru8274: Sorry, those are typos. In the actual code they were set right, but thanks for your answer.
David the H: Thanks, I will implement your suggestions. Thanks !