Re: RegEx Multi-Line Matching
Jeff Westman wrote:
On 6/20/05, John W. Krahn [EMAIL PROTECTED] wrote:
It looks like you don't really need to use paragraph mode, this should do what
you want:
#!/usr/bin/perl
#
# tnsnames.pl -- reads tnsnames.ora, sorts by host name
use warnings;
use strict;
my $tnsFile = 'tnsnames.ora';
open F, '', $tnsFile or die Could not open file $tnsFile: $!\n;
my ( $host, $name, %tns ) = ( '', '' );
while ( F ) {
$host = $1 if /\(HOST\s*=\s*(\w+)\)/;
$name = $1 if /\(SERVICE_NAME\s*=\s*(\w+)\)/;
if ( length $host and length $name ) {
print *** RESULT : host= $host, service= $name\n;
$tns{ $host } = $name;
( $host, $name ) = ( '', '' );
}
}
close F;
for ( sort keys %tns ) {
printf %-8s = %s\n, $_, $tns{ $_ };
}
__END__
Thank you for your post. I know I could have just done some
top-down scanning for pairs (I like your style, BTW), but it seemed
safer to do multi-line matching, so I am still a bit bothered why I
couldn't get the match to work in all cases (ie, a blank line was
mandatory).
I also had a basic question on the code you sent. You have:
$host = $1 if /\(HOST\s*=\s*(\w+)\)/;
$name = $1 if /\(SERVICE_NAME\s*=\s*(\w+)\)/;
This works and makes sense. But what if the line is commented?!
That is, when I modified this and made this:
$host = $1 if /[^#].*\(HOST\s*=\s*(\w+)\)/;
$name = $1 if /[^#].*\(SERVICE_NAME\s*=\s*(\w+)\)/;
it SHOULD have worked (I think!). It still matched a line that should
have been ignored. What am I doing wrong?!
The problem is that [^#] can match anywhere so in the string
ab#cd(HOST = xyz) it can match at 'a' or 'b' or 'c' or 'd'.
This should work better:
my $para = '';
while ( F ) {
$para .= $_;
next if $para =~ tr/(// != $para =~ tr/)//;
$para =~ s/#.*//g; # remove comment lines
if ( $para =~ /\(HOST\s*=\s*(\w+)\).*?\(SERVICE_NAME\s*=\s*(\w+)\)/s ) {
print *** RESULT : host= $1, service= $2\n;
$tns{ $1 } = $2;
}
$para = '';
}
for ( sort keys %tns ) {
printf %-8s = %s\n, $_, $tns{ $_ };
}
John
--
use Perl;
program
fulfillment
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/ http://learn.perl.org/first-response
Re: Regex Multi-Line Matching
Hey Chris, I'm not spamming anyone. I had an error when I sent my email(s) so I resent it. Sorry you had to click the mouse an additional time to 'delete'. Jeff On 6/20/05, Chris Devers [EMAIL PROTECTED] wrote: On Mon, 20 Jun 2005, Jeff Westman wrote: Any help would be greatly appreciated. We heard you the first two times. Please stop spamming the list with the same question. -- Chris Devers -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
RegEx Multi-Line Matching
Hello Fellow Perlites,
I'm having difficulty parsing a file and could use some help here.
I've attached my code, output file, and sample job run.
Basically, I want to read an Oracle TNS file (tnsnames.ora) and list
the host name and service on that server based on this input file. A
sample TNS entry is multi-lined and looke like this
mysprdtmp =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = mxr)(PORT = 1521))
)
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = mysprd01)
)
)
So, when the script runs, I would get this at end of job:
mx01 = tmsprd01
mx16 = mystst04
mxp = mysdev07
mxq = mystst02
mxr = mysprd01
s0869u03 = rdmtst06
ux08 = rdmtst03
uxb = scsdev01
The script generally works fine, but it has a problem when I do not
have a blank line in between these entries. In other words, if my
input file has
mysbse =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = mxp)(PORT = 1521))
)
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = mysbse)
)
)
oemdev =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = mxp)(PORT = 1521))
)
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = oemdev)
)
)
it should recognize the new TNS listing and give me
mxp = mysbse
mxp = oemdev
(along with the others I am parsing). Again, if I separate the two
with a blank line, it works fine. But since this is a DBA-owned file,
I cannot do this.
My regex looks like
my $tnsEntry = 'EOF';
^
\s*\w+\s*=\s*
\s+\(DESCRIPTION\s*=
\s+\(ADDRESS_LIST\s*=\s*
\s+\(ADDRESS\s*=\s*\(PROTOCOL\s*=\s*TCP\)\(HOST\s*=\s*(\w+)\)\(PORT\s*=\s*\d+\)\)
\s+\)\s*
\s+\(CONNECT_DATA\s*=\s*
\s+\(SERVER\s*=\s*\w+\)
\s+\(SERVICE_NAME\s*=\s*(\w+)\)
# \s+\)\s*
# \s+\)
$
EOF
Any help would be appreciated. I've been at this off and on for
several days. Note, this is not a production file, but just a
utility script.
Thanks
Jeff
--
perl script follows (tnsnames.pl):
--
#!/usr/bin/perl
#
# tnsnames.pl -- reads tnsnames.ora, sorts by host name
use warnings;
use strict;
my $host;
my $name;
my %tns;
my @sortKeys;
my $key;
my $val;
my $para;
my $matchNbr;
$/ = '';
# problem: tns entries with no line break are not distiquished
my $tnsEntry = 'EOF';
^
\s*\w+\s*=\s*
\s+\(DESCRIPTION\s*=
\s+\(ADDRESS_LIST\s*=\s*
\s+\(ADDRESS\s*=\s*\(PROTOCOL\s*=\s*TCP\)\(HOST\s*=\s*(\w+)\)\(PORT\s*=\s*\d+\)\)
\s+\)\s*
\s+\(CONNECT_DATA\s*=\s*
\s+\(SERVER\s*=\s*\w+\)
\s+\(SERVICE_NAME\s*=\s*(\w+)\)
# \s+\)\s*
# \s+\)
$
EOF
die ORACLE_HOME is not defined!\n unless (defined($ENV{ORACLE_HOME} ));
my $tnsFile = ./tnsnames.ora;
open(F, $tnsFile) or die Could not open file $tnsFile: $!\n;
while (chomp($para = F)) {
undef($host);
undef($name);
if ($para =~ m[$tnsEntry]msxo) {
$matchNbr++;
print -x80, \n;
print *** match #$matchNbr : $para\n;
$host = $1;
$name = $2;
print *** RESULT : host= $host, service= $name\n;
if ((defined($host)) (defined($name))) {
$tns{$host} = $name;
}
}
last if eof(F);
}
close(F);
@sortKeys = sort keys(%tns);
foreach (@sortKeys) {
printf %-8s = $tns{$_}\n, $_;
}
--
sample input file follows (tnsnames.ora)
--
# TNSNAMES.ORA Network Configuration File:
/u01/app/oracle/product/9.2.0/network/admin/tnsnames.ora
# Generated by Oracle configuration tools.
#
# history:
# 05/27/04,kk Added rdmdev02 and scsdev01
# 06/01/04,js Added rdmtst02 and rdmtst03 per j.jonawski
# 08/08/04,gv Added mysdev08
# 09/15/04,gv Added mystst04
# 09/20/04,cd Added mysdev04
## mxp ###
## Reports server
repserver_mysbse=(ADDRESS = (PROTOCOL = TCP)(HOST = mxp)(PORT = 1960))
repserver_mysdev01=(ADDRESS = (PROTOCOL = TCP)(HOST = mxp)(PORT = 1981))
repserver_mysdev02=(ADDRESS = (PROTOCOL = TCP)(HOST = mxp)(PORT = 1982))
repserver_mysdev03=(ADDRESS = (PROTOCOL = TCP)(HOST = mxp)(PORT = 1983))
repserver_mysdev04=(ADDRESS = (PROTOCOL = TCP)(HOST = mxp)(PORT = 1984))
repserver_mysdev05=(ADDRESS = (PROTOCOL = TCP)(HOST = mxp)(PORT = 1985))
repserver_mysdev06=(ADDRESS = (PROTOCOL = TCP)(HOST = mxp)(PORT = 1986))
repserver_mysdev07=(ADDRESS = (PROTOCOL = TCP)(HOST = mxp)(PORT = 1987))
repserver_mysdev08=(ADDRESS = (PROTOCOL = TCP)(HOST = mxp)(PORT = 1988))
repserver_mysdev04=(ADDRESS = (PROTOCOL = TCP)(HOST = mxp)(PORT = 1989))
## mxr ###
mysprdtmp =
(DESCRIPTION =
Regex Multi-Line Matching
My problem is simple. I want to parse an Oracle TNS file, listing
host and the service it provides.
A sample TNS entry is multi-lined and looks like this:
mysprdtmp =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = mxr)(PORT = 1521))
)
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = mysprd01)
)
)
The regex I am using to parse this looks like so:
mysprdtmp =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = mxr)(PORT = 1521))
)
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = mysprd01)
)
)
My code snippit is like this:
while (chomp($para = F)) {
undef($host);
undef($name);
if ($para =~ m[$tnsEntry]msxo) {
}
The problem is, that this works great, EXCEPT for cases where I do not
have a blank line separating two concatenated TNS entries (I only get
the last one).
Any help would be greatly appreciated.
Thanks
Jeff
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/ http://learn.perl.org/first-response
Re: Regex Multi-Line Matching
On Mon, 20 Jun 2005, Jeff Westman wrote: Any help would be greatly appreciated. We heard you the first two times. Please stop spamming the list with the same question. -- Chris Devers -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] http://learn.perl.org/ http://learn.perl.org/first-response
Re: RegEx Multi-Line Matching
Jeff Westman wrote:
Hello Fellow Perlites,
Hello,
I'm having difficulty parsing a file and could use some help here.
I've attached my code, output file, and sample job run.
Basically, I want to read an Oracle TNS file (tnsnames.ora) and list
the host name and service on that server based on this input file. A
sample TNS entry is multi-lined and looke like this
[snip]
The script generally works fine, but it has a problem when I do not
have a blank line in between these entries. In other words, if my
input file has
[snip]
(along with the others I am parsing). Again, if I separate the two
with a blank line, it works fine. But since this is a DBA-owned file,
I cannot do this.
[snip]
Any help would be appreciated. I've been at this off and on for
several days. Note, this is not a production file, but just a
utility script.
Thank you for providing a working program and data to test it with.
It looks like you don't really need to use paragraph mode, this should do what
you want:
#!/usr/bin/perl
#
# tnsnames.pl -- reads tnsnames.ora, sorts by host name
use warnings;
use strict;
my $tnsFile = 'tnsnames.ora';
open F, '', $tnsFile or die Could not open file $tnsFile: $!\n;
my ( $host, $name, %tns ) = ( '', '' );
while ( F ) {
$host = $1 if /\(HOST\s*=\s*(\w+)\)/;
$name = $1 if /\(SERVICE_NAME\s*=\s*(\w+)\)/;
if ( length $host and length $name ) {
print *** RESULT : host= $host, service= $name\n;
$tns{ $host } = $name;
( $host, $name ) = ( '', '' );
}
}
close F;
for ( sort keys %tns ) {
printf %-8s = %s\n, $_, $tns{ $_ };
}
__END__
John
--
use Perl;
program
fulfillment
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/ http://learn.perl.org/first-response
