Top Hosts computation should not use redirected urls and subframe urls

components/history/core/browser/history_database.cc

 // Copyright (c) 2012 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include "components/history/core/browser/history_database.h"
 
 #include <stdint.h>
 
 #include <algorithm>
 #include <set>
@@ skipping 175 matching lines @@
     UMA_HISTOGRAM_COUNTS_10000("History.WeeklyHostCount",
                                static_cast<int>(week_hosts.size()));
     UMA_HISTOGRAM_COUNTS("History.MonthlyURLCount", month_url_count);
     UMA_HISTOGRAM_COUNTS_10000("History.MonthlyHostCount",
                                static_cast<int>(month_hosts.size()));
     UMA_HISTOGRAM_TIMES("History.DatabaseAdvancedMetricsTime",
                         base::TimeTicks::Now() - start_time);
   }
 }
 
 TopHostsList HistoryDatabase::TopHosts(size_t num_hosts) {

[sky, 2016/08/31 16:29:47]

I'm not familiar with this function. I believe you had sdefresne review this
addition, please have him review the change.

   base::Time one_month_ago =
       std::max(base::Time::Now() - base::TimeDelta::FromDays(30), base::Time());
 
   sql::Statement url_sql(db_.GetUniqueStatement(
-      "SELECT url, visit_count FROM urls WHERE last_visit_time > ?"));
+      "SELECT u.url, u.visit_count "
+      "FROM urls u JOIN visits v ON u.id = v.url "
+      "WHERE last_visit_time > ? "
+      "AND (v.transition & ?) != 0 "              // CHAIN_END
+      "AND (transition & ?) NOT IN (?, ?, ?)"));  // NO SUBFRAME or
+                                                  // KEYWORD_GENERATED
+
   url_sql.BindInt64(0, one_month_ago.ToInternalValue());
+  url_sql.BindInt(1, ui::PAGE_TRANSITION_CHAIN_END);
+  url_sql.BindInt(2, ui::PAGE_TRANSITION_CORE_MASK);
+  url_sql.BindInt(3, ui::PAGE_TRANSITION_AUTO_SUBFRAME);
+  url_sql.BindInt(4, ui::PAGE_TRANSITION_MANUAL_SUBFRAME);
+  url_sql.BindInt(5, ui::PAGE_TRANSITION_KEYWORD_GENERATED);
 
   // Collect a map from host to visit count.
   base::hash_map<std::string, int> host_count;
   while (url_sql.Step()) {
     GURL url(url_sql.ColumnString(0));
     if (!(url.is_valid() && (url.SchemeIsHTTPOrHTTPS() || url.SchemeIs("ftp"))))
       continue;
 
     int64_t visit_count = url_sql.ColumnInt64(1);
     host_count[HostForTopHosts(url)] += visit_count;
@@ skipping 332 matching lines @@
       "SET visit_time = visit_time + 11644473600000000 "
       "WHERE id IN (SELECT id FROM visits WHERE visit_time > 0);"));
   ignore_result(db_.Execute(
       "UPDATE segment_usage "
       "SET time_slot = time_slot + 11644473600000000 "
       "WHERE id IN (SELECT id FROM segment_usage WHERE time_slot > 0);"));
 }
 #endif
 
 }  // namespace history