Guigou has asked for the wisdom of the Perl Monks concerning the following question:

Hi Monks,

I am very new to Perl coding. I want to achieve a simple taks which consists of getting starts and ends of sentences. All is stored in a file and looks like that :

<Turn speaker="spk2" startTime="6.97" endTime="8.536">

I had a meeting with a specialist who gave me a Perl code that looks like this : C:\_DATA_Guillaume_Nassau>cat C1_A1_1a.trs | perl -ne "print \"$1 $2 $2\n\" if (/Turn.*speaker=(.*)\s+.*=(.*)\s+.*=(.*)/)" > toto.xxx

That would compile a file with all the information I need. But it won't work --> "semicolon seem to be missing at line 1" Anyone sees the problem ? This worked perfectly on this person's computer, he copied and pasted it in an email sent to me. I did install activeperl which he was using and now I get this error...

Replies are listed 'Best First'.
Re: Help please?
by QM (Parson) on May 30, 2014 at 10:03 UTC
    Welcome to the monastery.

    For posting, use <code>code tags</code> around your code.

    Useless use of cat: Put the filename after all other perl arguments, such as: 

    perl -ne "blah" filename

    [download]

    You probably meant $3 instead of the duplicate $2.

    .* is greedy, and will tend to eat too much. Given your input example, you might use \S+ to only capture non-whitespace, and .*? to be non-greedy elsewhere. Or, perhaps constrain your regex more with /Turn.*?speaker=(\S+)\s+startTime=(\S+)\s+endTime=(\S+)/

    On Windoze/DOS, quoting is an issue, but you don't seem to have a problem so far. I prefer to use qq// and q// for quoting in perl oneliners. For example: 

    perl -ne "print qq/$1 $2 $3\n/ if (/Turn.*?speaker=(\S+)\s+startTime=(
+\S+)\s+endTime=(\S+)/)" input_file > toto.xxx

    [download]

    To check the compile without running, add -c as a perl option: 

    perl -c -ne "print qq/$1 $2 $3\n/ if (/Turn.*?speaker=(\S+)\s+startTim
+e=(\S+)\s+endTime=(\S+)/)"
-e syntax OK

    [download]

    -QM
    --
    Quantum Mechanics: The dreams stuff is made of

[reply]
[d/l]
[select]

      So I tried :

      perl -ne "print qq/$1 $2 $3\n/ if (/Turn.*?speaker=(\S+)\s+startTime=(+\S+)\s+endTime=(\S+)/)" input_file > toto.xxx

      And it tells me :

      c:\>perl -ne "print qq/$1 $2 $3\n/ if (/Turn.*?speaker=(\S+)\s+startTi
+me=(+\S+)\
s+endTime=(\S+)/)" C1_A1_1a.trs > toto.xxx
Quantifier follows nothing in regex; marked by <-- HERE in m/Turn.*?sp
+eaker=(\S+
)\s+startTime=(+ <-- HERE \S+)\s+endTime=(\S+)/ at -e line 1.

      [download]

      The file toto.xxx is created, but empty. I hardly understand what I am doing but I am in an effort to.

[reply]
[d/l]
[select]
        I observe that your copy/paste skills and/or typing skills have come up short. There's a stray + in (+\S+) (i.e., the first one.) The error message even points it out for you :)

        -QM
        --
        Quantum Mechanics: The dreams stuff is made of

[reply]
[d/l]
[select]
        How did the + get after the (?
        لսႽ† ᥲᥒ⚪⟊Ⴙᘓᖇ Ꮅᘓᖇ⎱ Ⴙᥲ𝇋ƙᘓᖇ
[reply]
[d/l]
Re: Help please?
by choroba (Cardinal) on May 30, 2014 at 09:49 UTC
    Did you copy&paste from the e-mail correctly? I fear you didn't, as the second $2 should've been $3. Isn't a backslash or quote missing somewhere?

    Also, your prompt seems MSWin-like. Are you sure cat works on MSWin?

    لսႽ† ᥲᥒ⚪⟊Ⴙᘓᖇ Ꮅᘓᖇ⎱ Ⴙᥲ𝇋ƙᘓᖇ
[reply]
[d/l]
      Sir, I am sure of absolutely nothing ahah. I did copy well because I used copy and paste so I guess it shouldn't add any error ? And the line was like that when that guy used it... I am not sure at all it does work on ms dos neither... might be the problem...
[reply]
        Outlook sometimes changes dash to emdash or somesuch. Go back and retype the dash, or perhaps retype the whole thing from scratch.

        -QM
        --
        Quantum Mechanics: The dreams stuff is made of

[reply]

Back to Seekers of Perl Wisdom