in reply to Re: splinting a line of text by comma
in thread splinting a line of text by comma

Rather than iterating through every permutation of every comma-separated sub-phrase, wouldn't it be better to print each set of sub-phrases once per line? Also, comma-separated sub-phrases at the beginning or end of the line will not be processed. How about something like:

c:\@Work\Perl\monks>perl -wMstrict -le "my @lines = ( 'some text foo | some text BAR | oh , no , commas | some text BAZ | + some text QUX', 'some text Oof | some, text, BAR | some , text , BAZ | some text QU +X', 'some, text, FOO | some text BAR | one two | some , text , QUX', ); ;; my $got_commas = qr{ (?: (?<= [|]) | \A) [^|,]* (?: , [^|,]*)+ (?= [|] | \z) }xms; ;; for my $line (@lines) { while ($line =~ m{ ($got_commas) }xmsg) { my $before = substr $line, 0, $-[1]; my $after = substr $line, $+[1]; print qq{:$before$_$after:} for split m{ , }xms, $1; } } " :some text foo | some text BAR | oh | some text BAZ | some text QUX: :some text foo | some text BAR | no | some text BAZ | some text QUX: :some text foo | some text BAR | commas | some text BAZ | some text QU +X: :some text Oof | some| some , text , BAZ | some text QUX: :some text Oof | text| some , text , BAZ | some text QUX: :some text Oof | BAR | some , text , BAZ | some text QUX: :some text Oof | some, text, BAR | some | some text QUX: :some text Oof | some, text, BAR | text | some text QUX: :some text Oof | some, text, BAR | BAZ | some text QUX: :some| some text BAR | one two | some , text , QUX: : text| some text BAR | one two | some , text , QUX: : FOO | some text BAR | one two | some , text , QUX: :some, text, FOO | some text BAR | one two | some : :some, text, FOO | some text BAR | one two | text : :some, text, FOO | some text BAR | one two | QUX:
(Availability of  \K with Perl versions 5.10+ can make the  (?: (?<= [|]) | \A)) expression a little nicer.)

Update: An "impure" variation. No big improvement; maybe of some interest.

c:\@Work\Perl\monks>perl -wMstrict -le "my @lines = ( 'some text foo | some text BAR | oh , no , commas | some text BAZ | + some text QUX', 'some text Oof | some, text, BAR | some , text , BAZ | some text QU +X', 'some, text, FOO | some text BAR | one two | some , text , QUX', ); ;; my $got_commas = qr{ (?: (?<= [|]) | \A) [^|,]* (?: , [^|,]*)+ (?= [|] | \z) }xms; ;; local our @out; use re 'eval'; m{ \A (.*?) ($got_commas) (?= (.*)) (?{ push @out, map qq{>$1$_$3<}, split m{ , }xms, $2 }) (?!) }xmsg for @lines; ;; print for @out; " >some text foo | some text BAR | oh | some text BAZ | some text QUX< >some text foo | some text BAR | no | some text BAZ | some text QUX< >some text foo | some text BAR | commas | some text BAZ | some text QU +X< >some text Oof | some| some , text , BAZ | some text QUX< >some text Oof | text| some , text , BAZ | some text QUX< >some text Oof | BAR | some , text , BAZ | some text QUX< >some text Oof | some, text, BAR | some | some text QUX< >some text Oof | some, text, BAR | text | some text QUX< >some text Oof | some, text, BAR | BAZ | some text QUX< >some| some text BAR | one two | some , text , QUX< > text| some text BAR | one two | some , text , QUX< > FOO | some text BAR | one two | some , text , QUX< >some, text, FOO | some text BAR | one two | some < >some, text, FOO | some text BAR | one two | text < >some, text, FOO | some text BAR | one two | QUX<


Give a man a fish:  <%-{-{-{-<

Replies are listed 'Best First'.
Re^3: splinting a line of text by comma
by tybalt89 (Monsignor) on Nov 23, 2018 at 02:14 UTC

    That is printing lines with commas. How about this with just a small tweak to allow commas in the first and/or last field.

    #!/usr/bin/perl # https://perlmonks.org/?node_id=1226189 use strict; use warnings; while( <DATA> ) { my @queue = $_; while( @queue ) { local $_ = shift @queue; if( / [^|,\n]* (?: , [^|,\n]* )+ /x ) { push @queue, map "$`$_$'", split /,/, $&; } else { print; } } } __DATA__ some text foo | some text BAR | oh , no , commas | some text BAZ | som +e text QUX some text Oof | some, text, BAR | some , text , BAZ | some text QUX some,text,FOO | some text BAR | one two | some , text , QUX

    Outputs:

    some text foo | some text BAR | oh | some text BAZ | some text QUX some text foo | some text BAR | no | some text BAZ | some text QUX some text foo | some text BAR | commas | some text BAZ | some text QUX some text Oof | some| some | some text QUX some text Oof | some| text | some text QUX some text Oof | some| BAZ | some text QUX some text Oof | text| some | some text QUX some text Oof | text| text | some text QUX some text Oof | text| BAZ | some text QUX some text Oof | BAR | some | some text QUX some text Oof | BAR | text | some text QUX some text Oof | BAR | BAZ | some text QUX some| some text BAR | one two | some some| some text BAR | one two | text some| some text BAR | one two | QUX text| some text BAR | one two | some text| some text BAR | one two | text text| some text BAR | one two | QUX FOO | some text BAR | one two | some FOO | some text BAR | one two | text FOO | some text BAR | one two | QUX
      ... one of the columns has multiple values separated with a comma like an array. ... unwind the array so each value of the array is on a separate line and all other fields are duplicated or carried with it ... [from the OP]

      Admittedly not the clearest specification. I was focusing on the "... all other fields are duplicated or carried with it ..." part, which I took to mean that some other fields with commas might be printed.

      ... mongo documentation of $unwind does what I'd like ...

      I don't understand what this means. Is "mongo" a thing?


      Give a man a fish:  <%-{-{-{-<

        I saw the "one of the columns" also and submitted my first solution. Then I thought "why not over design it", since the future is infinite and sometime in there someone will want multiple columns :)