Shell Programming and Scripting

Hi all,

I want to get each line of a data file from position 464 plus 8 characters. I tried in two different ways, and the results were different. I'd like to know why.

First method, using awk:

Code:

awk '{print substr($0,464,8)}' CONCIL_VUELTA_ALF_100112_0801.ok

Second method, using scripting:

Code:

while read line; do echo ${line: 463 :8}; done < CONCIL_VUELTA_ALF_100112_0801.ok

As far as I know, in awk position of columns starts at 1, that's why I started in position 464. However using shell expansion it starts at 0, so I started at 463. First of all, is that true?

Well, results are slightly different. Here's a little sample (diff command):
I don't know why, but awk sometimes fails because each line should have exactly 8 characters ('00000000'). While using scripting it works perfectly (<), in awk there are lines (>) that only have 6 or 7 characters.
Does anybody knows a reasonable explanation?

PD: I'm using cygwin 1.7.9(0.237/5/3) 2011-03-29 10:10

Thanks a lot.

Albert.

Hi Albert,
please post a sample of your input file.

Hi,

This is the first line that fails (awk only prints 7 characters, although awk output is 8 characters length, 7 zeros and a blank):
Whole line length is 546, but I don't know how post line with all blanks. Once I submit this post (or preview), blanks dissapear. So, I uploaded that line in attached sample1.txt

Other things I tried:

Code:

echo $text | awk '{print substr($0, 464, 8)}'

where $text is the line I posted above. It works, it prints 8 zeros.
I also tried on the file I'm uploading running

Code:

awk '{print substr($0, 464, 8)}' sample1.txt

It also works.
However when I ran the command over the whole file, it fails(which has more than 20 thousand lines and I cannot post because it has personal data):

Code:

awk '{print substr($0, 464, 8)}' CONCIL_VUELTA_ALF_100112_0801.ok

It prints:
It's an strange behaviour, because bold line is not the first occurrence. Blank lines are 100 characters length and I thought that could be the problem, but it cannot be. Sometimes awk prints right after 100 characters lines, sometimes it doesn't. I cannot find any pattern that explains why awk doesn't work properly and built-in functions does.

Thanks a lot, and sorry for my english xD

Albert.

We need a sample data to reproduce the issue ...
Could you try to modify or remove the sensitive data from the input file
and post it here as attachment?

Uff, it's too much work.
There are more than 20 thousand lines in that file, and not all of them has the same pattern.

I just wonder why similar commands had different output. I thought awk and scripting I posted before should work in the same way.
However, what I was attempting to do, it's done.

Thanks for your time.

Albert.

Depends on what you mean by similar ...
Consider the following:

Code:

zsh-4.3.14[t]% printf '\t 42\t \t \n'
         42              
zsh-4.3.14[t]% printf '\t 42\t \t \n' | awk '{ print length }'
8
zsh-4.3.14[t]% printf '\t 42\t \t \n' | { read; echo ${#REPLY};}
2
zsh-4.3.14[t]% printf '\t 42\t \t \n' | { IFS= read; echo ${#REPLY};}
8

And this is only part of the possible corner cases ...

Glad you've solved it.