Shell script for read user data with emptyLines in a text file and filter them

srimal · 10-29-2009, 06:12 AM

Hi all,
I have a file filled with user data following manner.The user details are arrange not in ordered manner and separated with an empty line.So it is needed to filter out them as below.

$cat myfile.txt

Code:

userid: ASDFG
userNo: 200834570000Z
SupName: ALSKFJDENC2
PostDetail: SUP_COLOMBO_INCIDENT_MAN|20080827|99991231|
PostDetail: SUP_KANDY_INCIDENT_MAN|20080827|99991231|
creator : ddddd

userid: KMVNBBCMX
SupName: KSJMNCBBXW3
creator: ssss
userNo: 209738270000Z
PostDetail: SUP_KEGALLE_INCIDENT_MAN|20080827|99991231|

userNo: 234235345358Z
SupName: MLAHSRTXVV4
userid: MLDKSURNTNVMXJ
PostDetail: SUP_KEGALLE_INCIDENT_MAN|20080827|99991231|

OutPut format:

Quote:

My Expected O/P:

Quote:

ASDFG|ALSKFJDENC2|200834570000Z|ddddd|SUP_COLOMBO_INCIDENT_MAN|SUP_KANDY_INCIDENT_MAN
KMVNBBCMX|KSJMNCBBXW3|209738270000Z|ssss|SUP_KEGALLE_INCIDENT_MAN
MLDKSURNTNVMXJ|MLAHSRTXVV4|234235345358Z||SUP_KEGALLE_INCIDENT_MAN

Can any one guide me to a solution please?
Thanks!

AngTheo789 · 10-29-2009, 07:39 AM

A shell script usually means something scripted in Bash or Ksh etc, but you can actually use many other scripting languages as well (like PHP, Perl or Python) and launch these on the command shell. You need to decide what scripting language you want to use - usually the one you know best.

catkin · 10-29-2009, 12:22 PM

Code:

#!/bin/bash
shopt -s extglob

PostDetail=''
while read line
do
    buf="${line%%:*}"
    keyword="${buf%% *}"
    buf="${line#*:}"
    data="${buf##*( )}"
    case "$keyword" in
        'userid' )
            userid="$data"
            ;;
        'userNo' )
            userNo="$data"
            ;;
        'SupName' )
            SupName="$data"
            ;;
        'PostDetail' )
            PostDetail="$PostDetail|${data%%|*}"
            ;;
        'creator' )
            creator="$data"
            ;;
        '' )
            echo "$userid|$SupName|$userNo|$creator$PostDetail"
            PostDetail=''
            ;;
    esac
done < input.txt
echo "$userid|$SupName|$userNo|$creator$PostDetail"

srimal · 10-30-2009, 05:51 AM

What is meant by

Code:

shopt -s extglob

I got an error like

Code:

shopt: not found

For out put, I got the last user has |ssss| entry instead of ||.
Any way Many Thanks for the guide. That'll be a great base for me to go ahead.

bartonski · 10-30-2009, 08:57 AM

Quote:

Originally Posted by AngTheo789

A shell script usually means something scripted in Bash or Ksh etc, but you can actually use many other scripting languages as well (like PHP, Perl or Python) and launch these on the command shell. You need to decide what scripting language you want to use - usually the one you know best.

All of the languages Theo mentioned have a data structure called a 'hash' or 'associative array', which is particularly well suited to this type of text manipulation. A hash is essentially an array which uses text fields to index the members of the array, rather than numbers. You lose the built in ordering of the array, but get an enormous amount of flexibility in return.

Here's my implementation in Perl:

Code:

#! /usr/bin/perl

sub printd {
    my @d;
    for $key ( qw(userid SupName userNo creator PostDetail) ){
        $h{$key} ||= "";
        push @d, $h{$key};
    }
    print join( '|', @d ) . "\n";
}

while(<>){
    chomp;
    if( ! /:/ ){
        printd;
        %h = (); # clear data hash after we've printed a section.
    } elsif ( /^PostDetail: ([^|]+)\|/ ) {
        $h{"PostDetail"} = defined $h{"PostDetail"} ? $h{"PostDetail"} . "|$1" : $1;
    } else {
        $_ =~ /^([^:]+): ?(.*)/;
        $h{$1} = $2;
    }
}
printd;

In this case '%h' is my hash. The key to the hash (ie the string which acts as an index), is the word before the colon, in each case.

The strength of this approach is that you don't have to create an enormous case statement to handle each one of your fields. With a limited data set such as yours, this is not a tremendous win, but if you had a few hundred fields, you would only have to expand the qw() expression within printd().

Note, for the sake of brevity, I've removed a number of perl's safety features such as 'use strict' and 'use warnings'. Don't do that in any code that you care about.

catkin · 10-31-2009, 11:05 AM

Quote:

Originally Posted by srimal

What is meant by

Code:

shopt -s extglob

I got an error like

Code:

shopt: not found

For out put, I got the last user has |ssss| entry instead of ||.
Any way Many Thanks for the guide. That'll be a great base for me to go ahead.

What is the first line of your script? #!/bin/bash or something else? Details of bash's shopt built-in are at the GNU Bash Reference Manual.

Here's a corrected version of the script

Code:

#!/bin/bash
shopt -s extglob

PostDetail=''
while read line
do
    buf="${line%%:*}"
    keyword="${buf%% *}"
    buf="${line#*:}"
    data="${buf##*( )}"
    case "$keyword" in
        'userid' )
            userid="$data"
            ;;
        'userNo' )
            userNo="$data"
            ;;
        'SupName' )
            SupName="$data"
            ;;
        'PostDetail' )
            PostDetail="$PostDetail|${data%%|*}"
            ;;
        'creator' )
            creator="$data"
            ;;
        '' )
            echo "$userid|$SupName|$userNo|$creator$PostDetail"
            PostDetail=''
            creator=''
            ;;
    esac
done < input.txt
echo "$userid|$SupName|$userNo|$creator$PostDetail"

bartonski · 10-31-2009, 04:43 PM

ps. Kudos to katkin; that is a fairly elegant piece of shell scripting. One thing that puzzles me: Just for kicks, I changed the hashbang line to '#! /bin/sh', and the script ran correctly. If I'm not mistaken, parameter expansion is not portable to posix, yet it works here... what's up with that?

catkin · 11-01-2009, 04:37 AM

Quote:

Originally Posted by bartonski

ps. Kudos to katkin; that is a fairly elegant piece of shell scripting. One thing that puzzles me: Just for kicks, I changed the hashbang line to '#! /bin/sh', and the script ran correctly. If I'm not mistaken, parameter expansion is not portable to posix, yet it works here... what's up with that?

Well, thank you bartonski, you are very kind

and you raise an interesting question ...

According to the GNU Bash Reference Manual's section on invoking bash, "If Bash is invoked with the name sh, it tries to mimic the startup behavior of historical versions of sh as closely as possible, while conforming to the posix standard as well". That's quite a juggling act because the historical Bourne Shell is a long way from POSIX compliance!

The GNU Bash Reference Manual's section on bash POSIX mode lists many differences between bash in native mode and bash in POSIX mode but it does not mention parameter expansion (except regards $PS1 and $PS2), shopt (except regards $PATH), extglob or filename expansion (except regards ~ and redirection).

Surprising? I was surprised

Why, then, does srimal's shell report "shopt: not found". Which shell is it? Will be interesting to learn what the first line of the script is.

BTW The shopt -s extglob is required for "${buf##*( )}"

Code:

c:~$ buf='    abc'
c:~$ echo "'${buf##*( )}'"
'    abc'
c:~$ shopt -s extglob
c:~$ echo "'${buf##*( )}'
'abc'