Fields script...

mmahulo · 01-15-2009, 07:36 AM

Hi

I am given a file that looks like this:

/dev/rhdisk5,00,00,CL2C,08:ed,OPEN-9-CVS-CM,100,10346,5003,0013,---,SMPL,SMPL,SMPL,SMPL,3,RAID5,10-16,L141F,L151F,L161F,L171F,XP,0x50060e8004286a02,00,7e,0x10800,U787A10000000C9466AAB:U787A10000000C9 3E95C3:,None,LVM,---,---,---,---,0x50060e8004286a02
/dev/rhdisk6,00,00,CL2C,00:37,OPEN-9,7042,10346,5003,0013,---,SMPL,SMPL,PVOL,SMPL,3,RAID5,10-16,L141F,L151F,L161F,L171F,XP,0x50060e8004286a02,00,7e,0x10800,U787A10000000C9466AAB:U787A10000000C9 3E95C3:,vgJRI2d,LVM,JRI2d_ora:/global/bc ,---,---,---,0x50060e8004286a02
/dev/rhdisk7,00,00,CL4C,01:37,OPEN-9,7042,10346,5003,0013,---,SMPL,SMPL,SMPL,SMPL,3,RAID5,10-16,L141F,L151F,L161F,L171F,XP,0x50060e8004286a22,00,7e,0x10900,U787A10000000C9466AAB:U787A10000000C9 3E95C3:,hpvg,LVM,tst1lv:/tst1 lv00:N/A ,---,---,---,0x50060e8004286a22
/dev/rhdisk8,00,00,CL2D,07:06,OPEN-9,7042,10346,5003,0013,---,SMPL,SMPL,SVOL,SMPL,3,RAID5,10-16,L141F,L151F,L161F,L171F,XP,0x50060e8004286a03,00,7e,0x10c00,U787A10000000C9466AAB:U787A10000000C9 3E95C3:,None,LVM,---,---,---,---,0x50060e8004286a03
/dev/rhdisk9,00,00,CL4D,07:3f,OPEN-9,7042,10346,5003,0013,---,SMPL,SMPL,SMPL,SMPL,3,RAID5,10-16,L141F,L151F,L161F,L171F,XP,0x50060e8004286a23,00,7e,0x10d00,U787A10000000C9466AAB:U787A10000000C9 3E95C3:,hbvg,LVM,,---,---,---,0x50060e8004286a23
/dev/rhdisk10,00,00,CL2B,00:1d,OPEN-9,7042,10320,5003,0013,---,SMPL,SMPL,PVOL,SMPL,3,RAID5,10-16,L141F,L151F,L161F,L171F,XP,0x50060e8004285001,00,7e,0x20400,U787A10000000C9466AAB:U787A10000000C9 3E95C3:,vgJRI2d,LVM,JRI2d_ora:/global/bc ,---,---,---,0x50060e8004285001
/dev/rhdisk11,00,00,CL4B,00:57,OPEN-9,7042,10320,5003,0013,---,SMPL,SMPL,PVOL,SMPL,3,RAID5,10-16,L141F,L151F,L161F,L171F,XP,0x50060e8004285021,00,7e,0x20500,U787A10000000C9466AAB:U787A10000000C9 3E95C3:,hpvg,LVM,tst1lv:/tst1 lv00:N/A ,---,---,---,0x50060e8004285021

From this file I need to find the following: Port (field 4), Device (field 1), Type (field 6), and Serial # (field 8).

The following is what I tried using arrays and each time I give it the array index to print a field associated with it (i.e. array index 0 for field 1, index 3 for field 4, etc) it prints the whole file for each of them.

my code is:

#!/bin/ksh
echo "Device, Serial#, Type, Port, CU:Ldev"
set -A Xpinfo
Xpinfo=`cat xpinfo_maker.out`
echo $Xpinfo[0], $Xpinfo[7], $Xpinfo[5], $Xpinfi[3], $Xpinfo[4]

Please help. Thanks in advance...

pixellany · 01-15-2009, 08:24 AM

When you post a code sample, it's best to tell us what it does (or does not) do.

Why not AWK? It is ideal for extracting specific fields.

mk27 · 01-15-2009, 08:31 AM

You have to parse that somehow. Since you are obviously just learning, I recommend you switch to perl since it's almost certainly better for this kind of task, is available everywhere, and will probably be more useful knowledge later.

Code:

#!/usr/bin/perl -w 
use strict; 

while (<DATA>) {
	$_=~/^\/dev\/(\w+),\d+,\d+,(\w+),\w+:\w+,([\w+-]+),\d+,(\d+),/;
	print "Device: $1 Port: $2 Type: $3 Serial Number: $4\n";
}

__DATA__
copy the data into here

I think the field numbers you give are fishing, but assuming you wanted the 1st, 4th, 6th, and 8th things seperated by commas, the output is:

Code:

Device: rhdisk5 Port: CL2C Type: OPEN-9-CVS-CM Serial Number: 10346
Device: rhdisk6 Port: CL2C Type: OPEN-9 Serial Number: 10346
Device: rhdisk7 Port: CL4C Type: OPEN-9 Serial Number: 10346
Device: rhdisk8 Port: CL2D Type: OPEN-9 Serial Number: 10346
Device: rhdisk9 Port: CL4D Type: OPEN-9 Serial Number: 10346
Device: rhdisk10 Port: CL2B Type: OPEN-9 Serial Number: 10320
Device: rhdisk11 Port: CL4B Type: OPEN-9 Serial Number: 10320

$1, $2, etc. correspond to substrings (in parantheses) in the =~/match/

I put the data at the end of the script for my own convenience, but you can read it from a file:

Code:

open (FH, "test.txt");

while (<FH>) {
	$_=~/^\/dev\/(\w+),\d+,\d+,(\w+),\w+:\w+,([\w+-]+),\d+,(\d+),/;
	print "Device: $1 Port: $2 Type: $3 Serial Number: $4\n";
}

close(FH);

jschiwal · 01-15-2009, 08:37 AM

I think it would take far less time learning to use awk to solve the problem then learning an entire language. Perl may be better at solving it, but learning to code in perl would take a very long time.

colucix · 01-15-2009, 09:13 AM

I agree to the aforementioned usage of awk to parse a file line by line.

Anyway, regarding your ksh script the problems are:
1) when assigning the content of the file to the array (using cat) it will split fields using blank spaces as delimiter (not comma)
2) the syntax to display a single array element is wrong. It should be echo ${Xpinfo[0]}

First consider the following script:

Code:

#!/bin/ksh
IFS=,
Xpinfo=( $(cat xpinfo_maker.out) )
echo ${#Xpinfo[@]}
echo ${Xpinfo[0]}, ${Xpinfo[7]}, ${Xpinfo[5]}, ${Xpinfo[3]}, ${Xpinfo[4]}

This uses comma as Input Field Separator so it assigns the content of the file to the array considering each comma separated field as an element. Look at the total count of elements stored in the array! Then it will print just the 0th, 7th, 5th, 3rd and 4th fields relative to the total content of the file. Conclusion: you have to parse the file line by line.

The following (maybe) is what you're looking for:

Code:

#!/bin/ksh
echo "Device, Serial#, Type, Port, CU:Ldev"

while read line
do
  echo $line | IFS=, read -A Xpinfo
  echo ${Xpinfo[0]}, ${Xpinfo[7]}, ${Xpinfo[5]}, ${Xpinfo[3]}, ${Xpinfo[4]}
done < xpinfo_maker.out

Also it is a good practice to save the original IFS at the beginning of the script - e.g. oldIFS=${IFS} - and then restore it when the modified IFS is not needed anymore - IFS=${oldIFS}.

mk27 · 01-15-2009, 09:49 AM

Quote:

Originally Posted by jschiwal

I think it would take far less time learning to use awk to solve the problem then learning an entire language. Perl may be better at solving it, but learning to code in perl would take a very long time.

Everytime you say this to yourself you will learn a little patch on something that will add up to a lot more if you just admit to yourself that you would be better off learning some perl. You don't have to learn it all and I guarantee learning to do this in perl will not be any more complicated than learning it in awk (which is also considered a language b/t/w)or with a shell, or whatever.

Anyway, the OP seems intent on learning shell scripting! So it looks like colucix saved the day.

pixellany · 01-15-2009, 03:32 PM

Quote:

admit to yourself that you would be better off learning some perl

Not knowing perl has not slowed me down one bit!!! YMMV

mmahulo · 01-16-2009, 12:33 AM

Quote:

Originally Posted by jschiwal

I think it would take far less time learning to use awk to solve the problem then learning an entire language. Perl may be better at solving it, but learning to code in perl would take a very long time.

Thank you for the post. Now, could you please help me solve this using AWK as havent really got the grips of it as much..?

pixellany · 01-16-2009, 08:28 AM

Here's a simple example:

awk -F, '{print $1, $3}' filename

This sets the field separator to "," and then prints the first and third columns

Really good tutorial here;
http://www.grymoire.com/Unix

mmahulo · 01-19-2009, 01:28 AM

Quote:

Originally Posted by pixellany

Here's a simple example:

awk -F, '{print $1, $3}' filename

This sets the field separator to "," and then prints the first and third columns

Really good tutorial here;
http://www.grymoire.com/Unix

Thanks pixellany. This works perfectly fine and it's fairly short. Thank you again