Re: replace/substituion 4th field

[7stud]

$line =~ s/(.*)\,(.*)\,(.*)\,(.*)\,(.*)\,(.*)/

Why are you escaping all those commas? A comma has no special regex meaning. In a regex, a comma means one thing: a literal comma.

The pattern .* says to match ANY character, 0 or more times, GREEDILY. That means that the regex engine will match as many characters as possible. When the regex engine sees a pattern like:

(.*),

...the first thing the regex engine does is find a match for the part in parentheses. Your whole line matches that part, so the regex engine reads in your whole line as the match for (.*). Then the regex engine moves on to the comma. Because there are no characters remaining in the line to match against, the regex engine backs up one character, surrendering a character from what matched .*. Then the regex engine checks if the comma matches that character. No match. So the regex engine backs up another character, surrendering yet another character from what matched .*, and the regex engine checks again if that character matches the comma. So on and so on until the regex finds a match for the comma. That's inefficient.

A more efficient regex would be this:

([^,]+),([^,]+)

However, as tmharish already mentioned you already split() your data, so there is no reason to use a regex at all. Just change whatever pieces you want in @fields like this:

$fileds[3] = "hurray";

You can limit your split to 5 instead of on every comma, which will speed things up a little.

It is possible to do a conditional s///, like this:

$line =~ s/
    (
        (?:
            [^,]+
            ,
        ){3}
    )
    ([^,]+)
    (.*)
/   $2 > 310    ? "$1volemd$3"
  : $2 == 70    ? "$1volemd1$3"
  : "$1$2$3"
  ;
/exms;

(The /e flag stands for eval.) But that is much slower than split() + join().

There is no reason to use Text::CSV_XS or any other CSV module.

[Anonymous Monk]

• There is no reason to use Text::CSV_XS or any other CSV module. For a csv file? Why is that?
• (The /e flag stands for eval.) No sir, the /e modifier does NOT means eval.
  This is what perlop said about it; A /e will cause the replacement portion to be treated as a full-fledged Perl expression and evaluated right then and there. It is, however, syntax checked at compile-time.
  It is when you have /ee, that the second "e" evaled, the expression. Please read it up.

[7stud]

There is no reason to use Text::CSV_XS or any other CSV module. For a csv file? Why is that?

Because perl natively supplies a simple and efficient tool to do what the op wants.

It is evaled when you have /ee, that the second "e" evaled, the expression. Please read it up.

I originally had two ee's, but that was so slow compared to split(), I changed it to one e.

It is, however, syntax checked at compile-time.

++ perl savoir-faire.