conditional statement in substitution expression

[newbio]

Hello Monks,

Example:

Input: **Type_1_deiodinase** , **D1** , metabolizes different forms, **A** , **B** of thyroid hormones to control levels of T3 , the active ligand for **thyroid_hormone_receptors** , **TR**

Output: **Type_1_deiodinase_(D1)** metabolizes different forms, **A_(B)** of thyroid hormones to control levels of T3 , the active ligand for **thyroid_hormone_receptors_(TR)**

If I do the following:  $line =~ s/\*\*([^\*]+)\*\*\s\,\s\*\*([^\*]+)\*\*(\s\,)?/**$1_($2)**/g;

it merges all $1 and $2 in the sentence. However, I want to merge terms only if either of $1 or $2 contains '_'.

That is in the above sentence, **Type_1_deiodinase_(D1)** and **thyroid_hormone_receptors_(TR)** are OK, while **A_(B)** is not. Is there a way to apply 'if' condition in the substitution expression above so that I can merge only those adjacent terms that contain '_'?

Thanks a lot.

[Roy Johnson]

  s/\*\*([^\*]+)\*\*\s\,\s\*\*([^\*]+)\*\*(\s\,)?/(grep {index($_,'_')
+>=0} $1,$2)?"**$1_($2)**":$&/ge;

This incurs the performance penalty for using $&, so you might prefer

while(<DATA>) {
  s/(\*\*([^\*]+)\*\*\s\,\s\*\*([^\*]+)\*\*(\s\,)?)/(grep {index($_,'_
+')>=0} $2,$3)?"**$2_($3)**":$1/ge; 
  print;
}

[jethro]

You might split the regex into two. In the first the regex should have the first ([^\*]+) changed to ([^\*]*_[^\*]*). In regex two the second ([^\*]+) is changed. Note this will also substitute if BOTH substrings contain underscores, but you could use ([^\*_]+) to prevent that.

[jwkrahn]

$line =~ s/\Q**\E(([^*]+)\Q**\E\s,\s\Q**\E([^*]+))\Q**\E(\s,)?/ "$2$3"
+ =~ tr!_!! ? "**$2_($3)**" : "**$1**" /eg;