in reply to Need Some help with finding a word in a file

If you know what is directly before the information you need, try changing the input record separator. Any time you think "unique", you most likely will want a hash.

#!/usr/bin/perl
use warnings;
use strict;

$/ = 'authDataAlias=';

my %no_dupes;

foreach my $line (<DATA>)
{

     if ($line =~ m/^"(.*?_DM\S+)"/i)
        {
        $no_dupes{$1} = 0;
        }
}

print "$_\n" for keys %no_dupes;

__DATA__
<factories xmi:type="resources.jdbc:CMPConnectorFactory" xmi:id="CMPCo
+nnectorFactory_1195273978412" name="dataSource" authMechanismPreferen
+ce="BASIC_PASSWORD" authDataAlias="cell-tstc-65_DM/userQ" connectionD
+efinition="ConnectionDefinition_1054132487569" cmpDatasource="DataSou
+rce_1195273954323"><factories xmi:type="resources.jdbc:CMPConnectorFa
+ctory" xmi:id="CMPConnectorFactory_1195273978412" name="dataSource" a
+uthMechanismPreference="BASIC_PASSWORD" authDataAlias="cell-tstc-65_D
+M/userQ" connectionDefinition="ConnectionDefinition_1054132487569" cm
+pDatasource="DataSource_1195273954323"><factories xmi:type="resources
+.jdbc:CMPConnectorFactory" xmi:id="CMPConnectorFactory_1195273978412"
+ name="dataSource" authMechanismPreference="BASIC_PASSWORD" authDataA
+lias="cell-tstc-65_DM/userF" connectionDefinition="ConnectionDefiniti
+on_1054132487569" cmpDatasource="DataSource_1195273954323">
<factories xmi:type="resources.jdbc:CMPConnectorFactory" xmi:id="CMPCo
+nnectorFactory_1195273978412" name="dataSource" authMechanismPreferen
+ce="BASIC_PASSWORD" authDataAlias="node-tstc-65_DM/userF" connectionD
+efinition="ConnectionDefinition_1054132487569" cmpDatasource="DataSou
+rce_1195273954323">

[download]

Replies are listed 'Best First'.
Re^2: Need Some help with finding a word in a file
by was6guy (Initiate) on Nov 29, 2007 at 21:33 UTC
    I think I need help with my regex, some of the stings have a null value, and some have a value, I need the ones that look like this:

    authDataAlias="cell-tstc-65_DM/userQ"

    I'm only concerned with: cell-tstc-65_DM/userQ

    If I do this, sed returns a blank line in the file since one of the authDataAlias strings is set to ="": 
    ($line =~ m/authDataAlias=\"([^\"]*)\"/i)

cell-tstc-65_DM/userQ

cell-tstc-65_DM/user1

    [download]
    If I run this sed command, it ignores the empty sting, but returns too much of the line: 
    ($line =~ m/authDataAlias=(.*-.*-.*_DM\/.*)/i)

"cell-tstc-65_DM/userQ" connectionDefinition="ConnectionDefinition_105
+4132487569" cmpDatasource="DataSource_1195273954323">
"cell-tstc-65_DM/user1" relationalResourceAdapter="builtin_rra" statem
+entCacheSize="10" datasourceHelperClassname="com.ibm.websphere.rsadap
+ter.DB2UniversalDataStoreHelper">

    [download]
[reply]
[d/l]
[select]
      Your above example works perfect. THANK YOU!
[reply]

In Section Seekers of Perl Wisdom