tange/tangetools
1
0
Fork 0
Code Issues Pull requests Actions Packages Projects Releases Wiki Activity

histogram/histogram: guessed delimiter must be present in all lines.

Browse source
This commit is contained in:
Ole Tange 2013-10-16 15:57:01 +02:00
parent f5e838293c
commit 69c72304d1
1 changed files with 15 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

17
histogram/histogram
View file

@ -449,7 +449,11 @@ for(my $i = 0; $i <= $#$value_ref; $i++) {
my $bar_length = $term_width - length($front_outside_string) - length($end_outside_string); my $bar_length = $term_width - length($front_outside_string) - length($end_outside_string);
my $factor; my $factor;
if($opt::log) { if($opt::log) {
if($value <= 0 or $max_value <= 0) {
$factor = 0;
} else {
$factor = log($value)/log($max_value); $factor = log($value)/log($max_value);
}
} else { } else {
$factor = $value/$max_value; $factor = $value/$max_value;
} }
@ -544,14 +548,23 @@ sub bar_string {
sub guess_delimiter { sub guess_delimiter {
my @raw = @_; my @raw = @_;
my %charcount; my (%charcount,$guess);
for(split//,join("",@raw)) { for(split//,join("",@raw)) {
# [a-zA-Z0-9] should never be auto chosen for delimiter # [a-zA-Z0-9] should never be auto chosen for delimiter
/[a-zA-Z0-9]/ and next; /[a-zA-Z0-9]/ and next;
$charcount{$_}++ $charcount{$_}++
} }
my $guess = (sort { $charcount{$b} <=> $charcount{$a} } keys %charcount)[0]; # The guess must be present in all lines
for my $g (sort { $charcount{$b} <=> $charcount{$a} } keys %charcount) {
defined $g or next;
if(grep { not /\Q$g\E/ } @raw) {
next;
} else {
$guess = $g;
last;
}
}
if(defined $guess and $guess =~ /\s/) { if(defined $guess and $guess =~ /\s/) {
# If the guess is a white space, then use 1+ whitespaces # If the guess is a white space, then use 1+ whitespaces
$guess = '\s+'; $guess = '\s+';