Merge branch 'language' into next

[usenet/newsstats.git] / gatherstats.pl

diff --git a/gatherstats.pl b/gatherstats.pl
index 2bc426b..6db137d 100755 (executable)
--- a/gatherstats.pl
+++ b/gatherstats.pl
@@ -1,4 +1,4 @@
-#! /usr/bin/perl -W
+#! /usr/bin/perl
 #
 # gatherstats.pl
 #
@@ -18,8 +18,9 @@ BEGIN {
   push(@INC, dirname($0));
 }
 use strict;
+use warnings;
 
-use NewsStats qw(:DEFAULT :TimePeriods ListNewsgroups ReadGroupList);
+use NewsStats qw(:DEFAULT :TimePeriods ListNewsgroups ParseHierarchies ReadGroupList);
 
 use DBI;
 use Getopt::Long qw(GetOptions);
@@ -132,7 +133,7 @@ foreach my $Month (&ListMonth($Period)) {
     # count postings per group
     my %Postings;
     while (($_) = $DBQuery->fetchrow_array) {
-      # get list oft newsgroups and hierarchies from Newsgroups:
+      # get list of newsgroups and hierarchies from Newsgroups:
       my %Newsgroups = ListNewsgroups($_,$TLH,
                                       $OptCheckgroupsFile ? \%ValidGroups : '');
       # count each newsgroup and hierarchy once
@@ -145,12 +146,21 @@ foreach my $Month (&ListMonth($Period)) {
     if (%ValidGroups) {
       foreach (sort keys %ValidGroups) {
         if (!defined($Postings{$_})) {
-          $Postings{$_} = 0 ;
+          # add current newsgroup as empty group
+          $Postings{$_} = 0;
           warn (sprintf("ADDED: %s as empty group\n",$_));
+          # add empty hierarchies for current newsgroup as needed
+          foreach (ParseHierarchies($_)) {
+            my $Hierarchy = $_ . '.ALL';
+            if (!defined($Postings{$Hierarchy})) {
+              $Postings{$Hierarchy} = 0;
+              warn (sprintf("ADDED: %s as empty group\n",$Hierarchy));
+            };
+          };
         }
       };
     };
-
+    
     # delete old data for that month
     if (!$OptTest) {
       $DBQuery = $DBHandle->do(sprintf("DELETE FROM %s.%s WHERE month = ?",
@@ -196,7 +206,7 @@ gatherstats - process statistical data from a raw source
 
 =head1 SYNOPSIS
 
-B<gatherstats> [B<-Vhdt>] [B<-m> I<YYYY-MM> | I<YYYY-MM:YYYY-MM>] [B<-s> I<stats] [B<-c> I<filename template>]] [B<--hierarchy> I<TLH>] [B<--rawdb> I<database table>] [B<-groupsdb> I<database table>] [B<--clientsdb> I<database table>] [B<--hostsdb> I<database table>]
+B<gatherstats> [B<-Vhdt>] [B<-m> I<YYYY-MM> | I<YYYY-MM:YYYY-MM>] [B<-s> I<stats>] [B<-c> I<filename template>]] [B<--hierarchy> I<TLH>] [B<--rawdb> I<database table>] [B<-groupsdb> I<database table>] [B<--clientsdb> I<database table>] [B<--hostsdb> I<database table>]
 
 =head1 REQUIREMENTS
 
@@ -283,7 +293,6 @@ Set processing period to a single month in YYYY-MM format or to a time
 period between two month in YYYY-MM:YYYY-MM format (two month, separated
 by a colon). 
 
-
 =item B<-s>, B<--stats> I<type>
 
 Set processing type to one of I<all> and I<groups>. Defaults to all
@@ -297,8 +306,9 @@ one group on each line and ignoring everything after the first
 whitespace (so you can use a file in checkgroups format or (part of)
 your INN active file).
 
-The filename is taken from I<filename template>, amended by each B<--
-month> B<gatherstats> is processing, so that
+The filename is taken from I<filename template>, amended by each
+B<--month> B<gatherstats> is processing in the form of I<template-YYYY-MM>,
+so that
 
     gatherstats -m 2010-01:2010-12 -c checkgroups