Perl regex to capture strings between anchor words - perl

Question

I am still working on cleaning up Oracle files, having to replace strings in files where the Oracle schema name is prepended to the function/procedure/package name within the file, as well as when the function/procedure/package name is double-quoted. Once the definition is corrected, I write the correction back to the file, along with the rest of the actual code.
I have code written to replace simple declarations (no input/output parameters) Now I am trying to get my regex to operate on (Note: This post is a continuation from this question) Some examples of what I'm trying to clean up:
Replace:
CREATE OR REPLACE FUNCTION "TRON2000"."DC_F_DUMP_CSV_MMA" (
p_trailing_separator IN BOOLEAN DEFAULT FALSE,
p_max_linesize IN NUMBER DEFAULT 32000,
p_mode IN VARCHAR2 DEFAULT 'w'
)
RETURN NUMBER
IS
to
CREATE OR REPLACE FUNCTION DC_F_DUMP_CSV_MMA (
p_trailing_separator IN BOOLEAN DEFAULT FALSE,
p_max_linesize IN NUMBER DEFAULT 32000,
p_mode IN VARCHAR2 DEFAULT 'w'
)
RETURN NUMBER
IS
I have been trying to use the following regex to separate the declaration, for later reconstruction after I've cleaned out the schema name / fixed the name of the function/procedure/package to not be double-quoted. I am struggling with getting each into a buffer - here's my latest attempt to grab all the middle input/output into it's own buffer:
\b(CREATE\sOR\sREPLACE\s(PACKAGE|PACKAGE\sBODY|PROCEDURE|FUNCTION))(?:\W+\w+){1,100}?\W+(RETURN)\s*(\W+\w+)\s(AS|IS)\b
Any / all help is GREATLY appreciated!
This is the script that I'm using right now to evaluate / write the corrected files:
#!/usr/bin/perl
use strict;
use warnings;
use File::Find;
use Data::Dumper;
# utility to clean strings
sub trim($) {
my $string = shift;
$string = "" if !defined($string);
$string =~ s/^\s+//;
$string =~ s/\s+$//;
# aggressive removal of blank lines
$string =~ s/\n+/\n/g;
return $string;
}
sub cleanup_packages {
my $file = shift;
my $tmp = $file . ".tmp";
my $package_name;
open( OLD, "< $file" ) or die "open $file: $!";
open( NEW, "> $tmp" ) or die "open $tmp: $!";
while ( my $line = <OLD> ) {
# look for the first line of the file to contain a CREATE OR REPLACE STATEMENT
if ( $line =~
m/^(CREATE\sOR\sREPLACE)\s*(PACKAGE|PACKAGE\sBODY)?\s(.+)\s(AS|IS)?/i
)
{
# look ahead to next line, in case the AS/IS is next
my $nextline = <OLD>;
# from the above IF clause, the package name is in buffer 3
$package_name = $3;
# if the package name and the AS/IS is on the same line, and
# the package name is quoted/prepended by the TRON2000 schema name
if ( $package_name =~ m/"TRON2000"\."(\w+)"(\s*|\S*)(AS|IS)/i ) {
# grab just the name and the AS/IS parts
$package_name =~ s/"TRON2000"\."(\w+)"(\s*|\S*)(AS|IS)/$1 $2/i;
trim($package_name);
}
elsif ( ( $package_name =~ m/"TRON2000"\."(\w+)"/i )
&& ( $nextline =~ m/(AS|IS)/ ) )
{
# if the AS/IS was on the next line from the name, put them together on one line
$package_name =~ s/"TRON2000"\."(\w+)"(\s*|\S*)/$1/i;
$package_name = trim($package_name) . ' ' . trim($nextline);
trim($package_name); # remove trailing carriage return
}
# now put the line back together
$line =~
s/^(CREATE\sOR\sREPLACE)\s*(PACKAGE|PACKAGE\sBODY|FUNCTION|PROCEDURE)?\s(.+)\s(AS|IS)?/$1 $2 $package_name/ig;
# and print it to the file
print NEW "$line\n";
}
else {
# just a normal line - print it to the temp file
print NEW $line or die "print $tmp: $!";
}
}
# close up the files
close(OLD) or die "close $file: $!";
close(NEW) or die "close $tmp: $!";
# rename the temp file as the original file name
unlink($file) or die "unlink $file: $!";
rename( $tmp, $file ) or die "can't rename $tmp to $file: $!";
}
# find and clean up oracle files
sub eachFile {
my $ext;
my $filename = $_;
my $fullpath = $File::Find::name;
if ( -f $filename ) {
($ext) = $filename =~ /(\.[^.]+)$/;
}
else {
# ignore non files
return;
}
if ( $ext =~ /(\.spp|\.sps|\.spb|\.sf|\.sp)/i ) {
print "package: $filename\n";
cleanup_packages($fullpath);
}
else {
print "$filename not specified for processing!\n";
}
}
MAIN:
{
my ( #files, $file );
my $dir = 'C:/1_atest';
# grab all the files for cleanup
find( \&eachFile, "$dir/" );
#open and evaluate each
foreach $file (#files)
{
# skip . and ..
next if ( $file =~ /^\.$/ );
next if ( $file =~ /^\.\.$/ );
cleanup_file($file);
};
}

Answer 1

Assuming the entire content of a file is stored as scalar in a var, the following should do the trick.
$Str = '
CREATE OR REPLACE FUNCTION "TRON2000"."DC_F_DUMP_CSV_MMA" (
p_trailing_separator IN BOOLEAN DEFAULT FALSE,
p_max_linesize IN NUMBER DEFAULT 32000,
p_mode IN VARCHAR2 DEFAULT w
)
RETURN NUMBER
IS
CREATE OR REPLACE FUNCTION "TRON2000"."DC_F_DUMP_CSV_MMA" (
p_trailing_separator IN BOOLEAN DEFAULT FALSE,
p_max_linesize IN NUMBER DEFAULT 32000,
p_mode IN VARCHAR2 DEFAULT w
)
RETURN NUMBER
IS
';
$Str =~ s#^(create\s+(?:or\s+replace\s+)?\w+\s+)"[^"]+"."([^"]+)"#$1 $2#mig;
print $Str;