perl parse across multiple line in txt file

i have a file with the following sample text

1 create table xyz
2 no before journal,
3 no after journal
4 (
5 col1 integer,
6 col2 integer,
7 ...
8 coln varchar(10)
9 )
10;
i want to use perl regex to search and replace text from the line that
starts with 'create' word till the first opening bracket i.e the '('
with blanks or rather delete the lines altogether.
the output should look something like this.


1 col1 integer,
2 col2 integer,
3 ...
4 coln varchar(10)
5 )
6;

the input file could also have the following scenarios..
1 create table xyz no before journal,
2 no after journal
3 (
4 col1 integer,
5 col2 integer,
6 ...
7 coln varchar(10)
8 )
9;

OR


1 create table xyz no before journal,
2 no after journal (
3 col1 integer,
4 col2 integer,
5 ...
6 coln varchar(10)
7 )
8;

OR

1 create table xyz
2 (
3 col1 integer,
4 col2 integer,
5 ...
6 coln varchar(10)
7 )
8;

the only certainty is that line starts with the 'create' word. THERE
COULD BE ANY WORDS BETWEEN THE 'CREATE' AND THE '(' .
so in short, i want the search to look for any line that begins with
the 'create' word and then continue the search till the first '(' and
replace the match with deleted lines.

I know how to use perl regex to search for one line at a time but not
if the condition could be across multiple lines.

any help will be greatly appreciated.

thanks

Re: perl parse across multiple line in txt file

bcdixit wrote:
> i have a file with the following sample text
>
> 1 create table xyz
> 2 no before journal,
> 3 no after journal
> 4 (
> 5 col1 integer,
> 6 col2 integer,
> 7 ...
> 8 coln varchar(10)
> 9 )
> 10;
> i want to use perl regex to search and replace text from the line that
> starts with 'create' word till the first opening bracket i.e the '('
> with blanks or rather delete the lines altogether.
> the output should look something like this.
>
>
> 1 col1 integer,
> 2 col2 integer,
> 3 ...
> 4 coln varchar(10)
> 5 )
> 6;


perldoc -q between

How can I pull out lines between two patterns that are themselves on dif‐
ferent lines?


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

my $cnt=1;
while ( ) {
next if /^\d+ create/ .. /^\d+ \(/;
s/^\d+/$cnt/;
$cnt++;
print;
}

__DATA__
1 create table xyz
2 no before journal,
3 no after journal
4 (
5 col1 integer,
6 col2 integer,
7 ...
8 coln varchar(10)
9 )
10;
--------------------------


--
Tad McClellan
email: perl -le "print scalar reverse qq/moc.noitatibaher\100cmdat/"

Re: perl parse across multiple line in txt file

bcdixit wrote:
> i have a file with the following sample text
>
> 1 create table xyz
> 2 no before journal,
> 3 no after journal
> 4 (
> 5 col1 integer,
> 6 col2 integer,
> 7 ...
> 8 coln varchar(10)
> 9 )
> 10;
> i want to use perl regex to search and replace text from the line that
> starts with 'create'


There is no line there that starts with 'create'.

There is a line that starts with '1', and with '2', and ...


--
Tad McClellan
email: perl -le "print scalar reverse qq/moc.noitatibaher\100cmdat/"

Re: perl parse across multiple line in txt file

sorry if I was not too clear. Those numbers are actually line numbers.
those numbers are not actually part of the text.

On Aug 18, 9:13 pm, Tad McClellan wrote:
> bcdixit wrote:
> > i have a file with the following sample text
>
> > 1 create table xyz
> > 2 no before journal,
> > 3 no after journal
> > 4 (
> > 5 col1 integer,
> > 6 col2 integer,
> > 7 ...
> > 8 coln varchar(10)
> > 9 )
> > 10;
> > i want to use perl regex to search and replace text from the line that
> > starts with 'create'
>
> There is no line there that starts with 'create'.
>
> There is a line that starts with '1', and with '2', and ...
>
> --
> Tad McClellan
> email: perl -le "print scalar reverse qq/moc.noitatibaher\100cmdat/"

Re: perl parse across multiple line in txt file

bcdixit wrote:
> i have a file with the following sample text
>
> 1 create table xyz
....
> 10;

Forget about lines, this is a case where you should be using ";" instead
of "\n" as the record delimiter. Check the docs for "input record delimiter".