Perl: perl regex for extracting values from complex lines

Question

Input log file:

 Nservdrx_cycle 4       servdrx4_cycle
 HCS_cellinfo_st[10]     (type = (LTE { 2}),cell_param_id = (28)
 freq_info =  (10560),band_ind = (rsrp_rsrq{ -1}),Qoffset1 = (0)
 Pcompensation = (0),Qrxlevmin = (-20),cell_id = (7), 
 agcreserved{3} = ({ 0, 0, 0 }))    
 channelisation_code1   16/5 { 4}   channelisation_code1
 sync_ul_info_st_   (availiable_sync_ul_code = (15),uppch_desired_power = 
 (20),power_ramping_step = (3),max_sync_ul_trans = (8),uppch_position_info =
 (0))
 trch_type  PCH { 7}    trch_type8      
 last_report    0   zeroth bit

I was trying to extract only integer for my above inputs but I am facing some issue with if the string contain integer at the beginning and at the end

For ( e.g agcreserved{3},HCS_cellinfo_st[10],Qoffset1) here I don't want to ignore {3},[10] and 1 but in my code it does. since I was extracting only integer.

Here I have written simple regex for extracting only integer.

MY SIMPLE CODE:

 use strict;
 use warnings;
 my $Ipfile  = 'data.txt';
 open my $FILE, "<", $Ipfile or die "Couldn't open input file: $!";
 my @array;
 while(<$FILE>)
 {
  while ($_ =~ m/( [+-]?\d+ )/xg)
  { 
   push @array, ($1);
  }

 }
print "@array \n";

output what I am getting for above inputs:

4 4 10 2 28 10560 -1 1 0 0 -20 7 3 0 0 0 1 16 5 4 1 15 20 3 8 0 7 8 0

expected output:

4 2 28 10560 -1 0 0 -20 7 0 0 0 4 15 20 3 8 0 7 0

If some body can help me with explanation ?

Notes. (1) If you match on $_ you don't need to write it -- while (/.../xg) will do. (Nothing really wrong with having $_ either.) (2) You want all matches in a string -- push @array, /.../xg;. The m/ (match operator) returns a list of matches. You don't have to go one by one. — zdim
– zdim, Commented Oct 26, 2016 at 16:35

ardavey · Accepted Answer · 2016-10-26 16:14:22Z

You are catching every integer because your regex has no restrictions on which characters can (or can not) come before/after the integer. Remember that the /x modifier only serves to allow whitespace/comments inside your pattern for readability.

Without knowing a bit more about the possible structure of your output data, this modification achieves the desired output:

  while ( $_ =~ m! [^[{/\w] ( [+-]?\d+ ) [^/\w]!xg ) {
    push @array, ($1);
  }

I have added rules before and after the integer to exclude certain characters. So now, we will only capture if:

There is no [, {, /, or word character immediately before the number
There is no / or word character immediately after the number

If your data could have 2-digit numbers in the { N} blocks (e.g. PCH {12}) then this will not capture those and the pattern will need to become much more complex. This solution is therefore quite brittle, without knowing more of the rules about your target data.

Collectives™ on Stack Overflow

Perl: perl regex for extracting values from complex lines

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related