tabdata/td-collapse

   1 #!/usr/bin/env perl
   2
   3 =pod
   4
   5 =head1 NAME
   6
   7 td-collapse - Collapse multiple tabular data records with equivalent keys into one.
   8
   9 =head1 SYNOPSIS
  10
  11 td-collapse [I<OPTIONS>]
  12
  13 =head1 DESCRIPTION
  14
  15 It goes row-by-row on a sorted tabular data stream
  16 and if 2 or more subsequent rows' first (key) cell are
  17 the same then collapse them into one row.
  18 This is done by joining corresponding cells' data from each row into one
  19 cell, effectively keeping every column's data in the same column.
  20
  21 If you want to group by an other column, not the first one, then first
  22 reorder the columns by td-select(1). Eg. C<td-select KEYCOLUMN +REST>.
  23
  24 =head1 OPTIONS
  25
  26 =over 4
  27
  28 =item -g, --glue I<STR>
  29
  30 Delimiter character or string between joined cell data.
  31 Default is space.
  32
  33 =back
  34
  35 =head1 EXAMPLES
  36
  37 This pipeline shows which users are using each of the configured default
  38 shells, grouped by shell path.
  39
  40   # get the list of users
  41   getent passwd |\
  42
  43   # transform into tabular data stream
  44   tr : "\t" |\
  45   td-add-headers USER X UID GID GECOS HOME SHELL |\
  46
  47   # put the shell in the first column, and sort, then collapse
  48   td-select SHELL USER | td-keepheader sort | td-collapse -g ' ' |\
  49
  50   # change header name "USER" to "USERS"
  51   td-alter USERS=USER | td-select +ALL -USER
  52
  53 B<Output>:
  54
  55   | COUNT | SHELL             | USERS                                        |
  56   | 4     | /bin/bash         | user1 user2 nova root                        |
  57   | 5     | /bin/false        | fetchmail hplip sddm speech-dispatcher sstpc |
  58   | 1     | /bin/sync         | sync                                         |
  59   | 1     | /sbin/rebootlogon | reboot                                       |
  60   | 6     | /usr/sbin/nologin | _apt avahi avahi-autoipd backup bin daemon   |
  61
  62 =head1 CAVEATS
  63
  64 Have to sort input data first.
  65
  66 =head1 SEE ALSO
  67
  68 td-expand(1) is a kind of inverse to td-collapse(1).
  69
  70 =head1 REFERENCES
  71
  72 td-collapse(1) roughly translates to SELECT COUNT(*) + GROUP_CONCAT() + GROUP BY in SQL.
  73
  74 =cut
  75
  76
  77 $OptGlue = " ";
  78 %OptionDefs = (
  79         'g|glue=s' => \$OptGlue,
  80 );
  81
  82 no if ($] >= 5.018), 'warnings' => 'experimental::smartmatch';
  83 do '/usr/lib/tool/perl5/tabdata/common.pl' or die "$@";
  84
  85 process_header(scalar <STDIN>);
  86 unshift @Header, "COUNT";
  87 print join($FS, @Header).$RS;
  88
  89
  90 sub output_last_group
  91 {
  92         print $collapsed_rows . $FS . $prev_group_key . $FS . join($FS, map { join $OptGlue, @$_ } @group_members) . $RS;
  93         @group_members = ();
  94         $collapsed_rows = 0;
  95 }
  96
  97 $group_key = undef;
  98 $prev_group_key = undef;
  99 @group_members = ();
 100 $collapsed_rows = 0;
 101
 102 while(not eof STDIN)
 103 {
 104         my @input_row = read_record(\*STDIN);
 105
 106         $group_key = $input_row[0];
 107         output_last_group if defined $prev_group_key and $group_key ne $prev_group_key;
 108         for my $cell_num (1..$#input_row)
 109         {
 110                 push @{$group_members[$cell_num-1]}, $input_row[$cell_num];
 111         }
 112         $collapsed_rows++;
 113         $prev_group_key = $group_key;
 114 }
 115
 116 END { output_last_group; }