Figuring out the regex problem

[Ramnarayan]

Hi,

I have completely written the script to find the _it, _ft, _wt, _wa, _ua, _ba, _wk, _yc, _1g to _9g fields from the below list. Now these fields can appear only within a _t3 field which is given below. Hence, if a _t3 field contains atleast one of the (_it, _ft, _wt, _wa, _ua, _ba, _wk, yc, _1g to _9g), It should add 1 to $trans variable.

Here is the snippet of the code:

open (DTC, "<$toc") or fatal("Can't open $toc: $!");
while (my $line = <DTC>)
{
if ($line =~ m/^(_t3\s*.+\r?\n)$/)
{
$flag = 1;
}
else {$flag = 0;}
if ($line =~ m/^_(it|ft|wt|wa|ua|ba|wk|yc|[1-9]g)/ && $flag == 1)
{
++$trans;
}
}
close DTC;
print "Trans: $trans\n";

Here is the extract of the toc file. This has examples of _it and _wt

_t3 AP000024 00368075 AP990837 99A00100
_g1 Scientific Books
_ty BRV
_tw Reproduction artificielle de mineraux au XIXe siecle
_wt Reproduction artificielle de minéraux au XIXe siècle
_aw P. N. Tchirwinsky
_pg 66-68
_mf [Raw ASCII] 26 27 28
_mf [TIFF 6.0] 26 27 28
_t3 AP000024 00368075 AP990837 99A00110
_ti Miastor Larvae
_it Miastor Larvæ
_au E. P. Felt
_pg 583
_mf [Raw ASCII] 31
_mf [TIFF 6.0] 31

==================
Please note that each t3 may contain any of the (_it, _ft, _wt, _wa, _ua, _ba, _wk, yc, _1g to _9g). Either of these fields can occur anywhere within a _t3 block.

Can someone help me out. I will be very grateful to whoever can help me here!

Thanks

 

[uida1154]

Sorry but can you please rephrase the question? What should the script do?

 

[mikevh]

I agree with uida1154. You need to explain more clearly what you're trying to do. For example, what is your definition of a "field"? Is it the same as a line of data in your example? What do you mean when you say
"these fields can appear only within a _t3 field"?

 

[Ramnarayan]

The "field" corresponds to either of the fields in the below list:
_it, _ft, _wt, _wa, _ua, _ba, _wk, yc, _1g to _9g


So if there is a _t3 with a _it, _ft as below:

_t3 Blah
_it blah blah
_ft blah blah blah

It should cound one occurence to the variable $trans.

Hope this makes sense!

 

[mikevh]

Where does a t3 field end? Does it end when we see a line that begins with something that is not
_it, _ft, _wt, _wa, _ua, _ba, _wk, yc, _1g to _9g?

For example, from your data above:
_t3 AP000024 00368075 AP990837 99A00100
_g1 Scientific Books <- t3 field ends here?

_t3 AP000024 00368075 AP990837 99A00110
_ti Miastor Larvae
_it Miastor Larvæ
_au E. P. Felt <- t3 field ends here?

So, using your example data, $trans would equal 2 at the end of the loop?

 

[Ramnarayan]

Yes that is exactly right! To make it simple, the _t3 ends with another _t3. That is all.

 

[mikevh]

Okay, so once we've seen a t3, it never gets turned off, since the next one we see turns it back on?
I think this will do it:

while (my $line = <DTC>)
{
    if ($line =~ m/^_t3/)
    {
           $flag = 1;
           next;
    }
    if ($line =~ m/^_(it|ft|wt|wa|ua|ba|wk|yc|[1-9]g)/)
    {
           if ($flag)
           {
              ++$trans;
           }
    }
}
print "Trans: $trans\n";

This prints 2 with your example data.

 

[laserBeam]

Assuming the regex works....
Since He wants to count 1 occurrence of the match between _t3, you need to reset the $flag to 0 where you are incrementing.

if ($flag)
{
++$trans;
$flag = 0;
}

 

[Ramnarayan]

Now it works fine. Thanks for your valuable time and patience. I appreciate your efforts to help me out.
Laserbeam: Your resetting of $flag = 0 is correct. If that was not there, it will gives a wrong count.