null

gromlok · 2011-09-25 01:38:16

null

Last edited by gromlok (2019-03-10 15:45:23)

igndenok · 2011-09-25 01:58:36

I) Check this [ vs. [[
II) IIRC, you need something like this in your input file

string1\ string2\ string3

Instead this, because $line only hold 1st column

string1 string2 string3

Even after using '\ ' use "$line" instead $line.

PS: I didn't check the actual error, but this is from my personal experience when using `while read ...' (assuming there's space between strings in your string list)
III) Maybe using something like this

counter=$(expr $counter + 1)

Last edited by igndenok (2011-09-25 02:00:51)

falconindy · 2011-09-25 02:56:47

This is a rather suitable job for awk....

awk 'length($0) > 100 && $0 ~ /^5698/ { cnt++; print } END { printf "Number of Register: %d\n", cnt > "/dev/stderr" }' inputfile > outputfile

Other things to note about your shell implementation:

- You cannot invoke a bash script with 'sh' and expect it to be run as a bash script. Even when /bin/sh links to /bin/bash, the behavior of the shell is different (more POSIX like).
- Your redirection ($line > output.txt) doesn't work because you're trying to execute a command, not write to a file. use echo or printf, and please quote properly.

If you wanted this done in shell, its a little more verbose than the awk...

#!/bin/bash

exec 3>outputfile

while read -r line; do 
  (( ${#line} > 100 )) && [[ ${line:0:4} = 5968 ]] || continue
  (( ++count ))
  printf '%s\n' "$line" >>&3
done < inputfile

printf 'Number of registers: %s\n' "$count"

I'll just leave this here. It's far more suitable for beginners than TLDP, mostly because its factually correct and gives several examples of what not to do, which is just as important when writing shell.

Last edited by falconindy (2011-09-25 03:53:58)

gromlok · 2011-09-25 03:51:57

null

Last edited by gromlok (2019-03-10 15:45:38)

falconindy · 2011-09-25 04:01:18

gromlok wrote:

I) In your code, why use:
 done < inputfile 
. I don't get it. With the "while" is not enough to read all the file.

Not really sure I understand this... 'read' in a while loop will read from stdin until it reaches EOF, or encounters a read error. The while loop as I wrote it will read the input file in its entirety as the file is replacing standard input for the duration of the loop.

gromlok wrote:

II) Is it OK to use inputfile=$1 and outputfile=$2 that way?

Sure. Different variable name, same value.

gromlok wrote:

III) I did not find why (( ++count )) works in shell! Using
 let count = count + 1 
works, but that way you did it is more elegant.

If (( ++count )) doesn't work, then you're still executing the script with /bin/sh.

gromlok wrote:

IV) Puting "exit" at the end is redundant?

Entirely redundant if you're not adding an argument.

Stebalien · 2011-09-25 04:59:44

You can also use:

let count++

gromlok · 2011-09-25 16:39:15

null

Last edited by gromlok (2019-03-10 15:45:49)

alphaniner · 2011-09-29 20:38:20

falconindy wrote:

I'll just leave this here...

Absolutely one of the best resources on bash I've ever come across. Of course, after finding it I was forced to do a major overhaul on all of my scripts.

Arch Linux

#1 2011-09-25 01:38:16

null

#2 2011-09-25 01:58:36

Re: null

#3 2011-09-25 02:56:47

Re: null

#4 2011-09-25 03:51:57

Re: null

#5 2011-09-25 04:01:18

Re: null

#6 2011-09-25 04:59:44

Re: null

#7 2011-09-25 16:39:15

Re: null

#8 2011-09-29 20:38:20

Re: null

Board footer