Ytukyg

I have many sentences:

 1) the 3d line chart will show area in 3d.

 2) udcv123hi2ry32 the this line chart is useful.

 3) this chart.

 4) a chart.

 5) a line chart.

 6) this bar chart

 7) ...

And I have conditions

 1) substrings start by 'a' or 'the' or 'this' or '[chart name]'

 2) '[chart name] chart' is ok but 'this chart', 'a chart' are not accepted.

    (e.g. bar chart, line chart, this line chart, a area chart: OK,

     this chart, a chart: not accepted)

 3) substrings end by '.(dot)'

consequently, i need to find substrings that meet condition.

In this case the string:

"this line chart is very useful.", 

"area chart is very useful." are exactly what I want to receive.

I try to this via regular expression like this(https://regex101.com/r/aX5htr/2):

(a|the|this)* *((?!bthis chartb|bwhich chartb|ba chartb|bthe chartb|bthat chartb|d+).+ chart) .+.

but not matched...

how to solve this situations??

You may use

my $rx = qr/(?x)                 # enable formatting whitespace/comments

    (?(DEFINE)                   # Start DEFINE block

      (?<start>a|the|this|which) # Match start delimiters

    )                            # End DEFINE block

    (?<res>                      # Group res holding the match

      b(?&start)s+chartb      # Match start delims, 1+ whitespace, chart

      (*SKIP)(*F)                # and skip the match

      |                          # or

      b(?:(?&start)s+)?        # Optional start delim and 1+ whitespace

      w+s+chartb              # 1+ word chars, 1+ whitespace, char, word boundary

      [^.]*                      # 0+ chars other than dot

    )                            # End of res group

/;

See the regex demo.

See the Perl demo online:

use strict;

use warnings;



my $rx = qr/(?x)                 # enable formatting whitespace/comments

    (?(DEFINE)                   # Start DEFINE block

      (?<start>a|the|this|which) # Match start delimiters

    )                            # End DEFINE block

    (?<res>                      # Group res holding the match

      b(?&start)s+chartb      # Match start delims, 1+ whitespace, chart

      (*SKIP)(*F)                # and skip the match

      |                          # or

      b(?:(?&start)s+)?        # Optional start delim and 1+ whitespace

      w+s+chartb              # 1+ word chars, 1+ whitespace, char, word boundary

      [^.]*                      # 0+ chars other than dot

    )                            # End of res group

/;

while (<DATA>) {

    if (/$rx/) {

        print "$+{res}n";

    }

}



__DATA__

this chart.

this line chart.

this bar chart.

21684564523 this chart.

556465465456 this a line chart.

a chart.

a line chart.

which chart.

all this chart.

a chart.

123123 this chart..

123123 which chart.

all this line chart.

a line chart.

the 3d line chart will show area in 3d.

line chart.

area chart.

the chart.

1221513513 line chart.

1234125135 the chart.

123123 this bar chart.

udcvhi2ry32 the this line chart is useful.

twl chart.

Output:

this line chart

this bar chart

a line chart

a line chart

this line chart

a line chart

line chart will show area in 3d

line chart

area chart

line chart

this bar chart

this line chart is useful

twl chart