Merge branch 'master' into fix-netloc

.gitignore

@@ -79,6 +79,7 @@ configure-stamp
 *.bin
 *.mrk
 *.mrk2
+*.mrk3
 
 .dupload.conf
 

README.md

@@ -10,6 +10,7 @@ ClickHouse is an open-source column-oriented database management system that all
 * [YouTube channel](https://www.youtube.com/c/ClickHouseDB) has a lot of content about ClickHouse in video format.
 * [Slack](https://join.slack.com/t/clickhousedb/shared_invite/zt-d2zxkf9e-XyxDa_ucfPxzuH4SJIm~Ng) and [Telegram](https://telegram.me/clickhouse_en) allow to chat with ClickHouse users in real-time.
 * [Blog](https://clickhouse.yandex/blog/en/) contains various ClickHouse-related articles, as well as announcements and reports about events.
+* [Code Browser](https://clickhouse.tech/codebrowser/html_report/ClickHouse/index.html) with syntax highlight and navigation.
 * [Yandex.Messenger channel](https://yandex.ru/chat/#/join/20e380d9-c7be-4123-ab06-e95fb946975e) shares announcements and useful links in Russian.
 * [Contacts](https://clickhouse.tech/#contacts) can help to get your questions answered if there are any.
 * You can also [fill this form](https://clickhouse.tech/#meet) to meet Yandex ClickHouse team in person.

cmake/autogenerated_versions.txt

@@ -1,9 +1,9 @@
 # This strings autochanged from release_lib.sh:
-SET(VERSION_REVISION 54437)
+SET(VERSION_REVISION 54438)
 SET(VERSION_MAJOR 20)
-SET(VERSION_MINOR 7)
+SET(VERSION_MINOR 8)
 SET(VERSION_PATCH 1)
-SET(VERSION_GITHASH d64e51d1a78c1b53c33915ca0f75c97b2333844f)
+SET(VERSION_GITHASH 5d60ab33a511efd149c7c3de77c0dd4b81e65b13)
-SET(VERSION_DESCRIBE v20.7.1.1-prestable)
+SET(VERSION_DESCRIBE v20.8.1.1-prestable)
-SET(VERSION_STRING 20.7.1.1)
+SET(VERSION_STRING 20.8.1.1)
 # end of autochange

debian/changelog

@@ -1,5 +1,5 @@
-clickhouse (20.7.1.1) unstable; urgency=low
+clickhouse (20.8.1.1) unstable; urgency=low
 
   * Modified source code
 
- -- clickhouse-release <clickhouse-release@yandex-team.ru>  Mon, 13 Jul 2020 18:25:58 +0300
+ -- clickhouse-release <clickhouse-release@yandex-team.ru>  Fri, 07 Aug 2020 21:45:46 +0300

docker/client/Dockerfile

@@ -1,7 +1,7 @@
 FROM ubuntu:18.04
 
 ARG repository="deb https://repo.clickhouse.tech/deb/stable/ main/"
-ARG version=20.7.1.*
+ARG version=20.8.1.*
 
 RUN apt-get update \
     && apt-get install --yes --no-install-recommends \

docker/packager/deb/Dockerfile

@@ -21,7 +21,7 @@ RUN apt-get --allow-unauthenticated update -y \
 # Special dpkg-deb (https://github.com/ClickHouse-Extras/dpkg) version which is able
 # to compress files using pigz (https://zlib.net/pigz/) instead of gzip.
 # Significantly increase deb packaging speed and compatible with old systems
-RUN curl -O https://clickhouse-builds.s3.yandex.net/utils/dpkg-deb
+RUN curl -O https://clickhouse-builds.s3.yandex.net/utils/1/dpkg-deb
 RUN chmod +x dpkg-deb
 RUN cp dpkg-deb /usr/bin
 

docker/server/Dockerfile

@@ -1,7 +1,7 @@
 FROM ubuntu:20.04
 
 ARG repository="deb https://repo.clickhouse.tech/deb/stable/ main/"
-ARG version=20.7.1.*
+ARG version=20.8.1.*
 ARG gosu_ver=1.10
 
 RUN apt-get update \

docker/test/Dockerfile

@@ -1,7 +1,7 @@
 FROM ubuntu:18.04
 
 ARG repository="deb https://repo.clickhouse.tech/deb/stable/ main/"
-ARG version=20.7.1.*
+ARG version=20.8.1.*
 
 RUN apt-get update && \
     apt-get install -y apt-transport-https dirmngr && \

docker/test/performance-comparison/report.py

@@ -179,10 +179,10 @@ def advanceRowAnchor():
     return currentRowAnchor()
 
 
-def tr(x):
+def tr(x, anchor=None):
-    a = advanceRowAnchor()
     #return '<tr onclick="location.href=\'#{a}\'" id={a}>{x}</tr>'.format(a=a, x=str(x))
-    return '<tr id={a}>{x}</tr>'.format(a=a, x=str(x))
+    anchor = anchor if anchor else advanceRowAnchor()
+    return f'<tr id={anchor}>{x}</tr>'
 
 def td(value, cell_attributes = ''):
     return '<td {cell_attributes}>{value}</td>'.format(
@@ -192,12 +192,14 @@ def td(value, cell_attributes = ''):
 def th(x):
     return '<th>' + str(x) + '</th>'
 
-def tableRow(cell_values, cell_attributes = []):
+def tableRow(cell_values, cell_attributes = [], anchor=None):
-    return tr(''.join([td(v, a)
+    return tr(
-        for v, a in itertools.zip_longest(
+        ''.join([td(v, a)
-            cell_values, cell_attributes,
+            for v, a in itertools.zip_longest(
-            fillvalue = '')
+                cell_values, cell_attributes,
-        if a is not None and v is not None]))
+                fillvalue = '')
+            if a is not None and v is not None]),
+        anchor)
 
 def tableHeader(r):
     return tr(''.join([th(f) for f in r]))
@@ -291,8 +293,8 @@ def add_errors_explained():
     if not errors_explained:
         return
 
-    text = tableStart('Error summary')
+    text = '<a name="fail1"/>'
-    text += '<a name="fail1"/>'
+    text += tableStart('Error summary')
     text += tableHeader(['Description'])
     for row in errors_explained:
         text += tableRow(row)
@@ -342,19 +344,20 @@ if args.report == 'main':
         text += tableHeader(columns)
         attrs = ['' for c in columns]
         for row in rows:
+            anchor = f'{currentTableAnchor()}.{row[2]}.{row[3]}'
             if float(row[1]) > 0.10:
                 attrs[1] = f'style="background: {color_bad}"'
                 unstable_partial_queries += 1
-                errors_explained.append([f'<a href="#{nextRowAnchor()}">The query no. {row[3]} of test \'{row[2]}\' has excessive variance of run time. Keep it below 10%</a>'])
+                errors_explained.append([f'<a href="#{anchor}">The query no. {row[3]} of test \'{row[2]}\' has excessive variance of run time. Keep it below 10%</a>'])
             else:
                 attrs[1] = ''
             if float(row[0]) > allowed_single_run_time:
                 attrs[0] = f'style="background: {color_bad}"'
-                errors_explained.append([f'<a href="#{nextRowAnchor()}">The query no. {row[3]} of test \'{row[2]}\' is taking too long to run. Keep the run time below {allowed_single_run} seconds"</a>'])
+                errors_explained.append([f'<a href="#{anchor}">The query no. {row[3]} of test \'{row[2]}\' is taking too long to run. Keep the run time below {allowed_single_run} seconds"</a>'])
                 slow_average_tests += 1
             else:
                 attrs[0] = ''
-            text += tableRow(row, attrs)
+            text += tableRow(row, attrs, anchor)
         text += tableEnd()
         tables.append(text)
 
@@ -385,6 +388,7 @@ if args.report == 'main':
         attrs = ['' for c in columns]
         attrs[5] = None
         for row in rows:
+            anchor = f'{currentTableAnchor()}.{row[6]}.{row[7]}'
             if int(row[5]):
                 if float(row[3]) < 0.:
                     faster_queries += 1
@@ -392,11 +396,11 @@ if args.report == 'main':
                 else:
                     slower_queries += 1
                     attrs[2] = attrs[3] = f'style="background: {color_bad}"'
-                    errors_explained.append([f'<a href="#{nextRowAnchor()}">The query no. {row[7]} of test \'{row[6]}\' has slowed down</a>'])
+                    errors_explained.append([f'<a href="#{anchor}">The query no. {row[7]} of test \'{row[6]}\' has slowed down</a>'])
             else:
                 attrs[2] = attrs[3] = ''
 
-            text += tableRow(row, attrs)
+            text += tableRow(row, attrs, anchor)
 
         text += tableEnd()
         tables.append(text)
@@ -429,13 +433,14 @@ if args.report == 'main':
         attrs = ['' for c in columns]
         attrs[4] = None
         for r in unstable_rows:
+            anchor = f'{currentTableAnchor()}.{r[5]}.{r[6]}'
             if int(r[4]):
                 very_unstable_queries += 1
                 attrs[3] = f'style="background: {color_bad}"'
             else:
                 attrs[3] = ''
 
-            text += tableRow(r, attrs)
+            text += tableRow(r, attrs, anchor)
 
         text += tableEnd()
         tables.append(text)
@@ -477,14 +482,14 @@ if args.report == 'main':
                 # FIXME should be 15s max -- investigate parallel_insert
                 slow_average_tests += 1
                 attrs[6] = f'style="background: {color_bad}"'
-                errors_explained.append([f'<a href="./all-queries.html#all-query-times.0">The test \'{r[0]}\' is too slow to run as a whole. Investigate whether the create and fill queries can be sped up'])
+                errors_explained.append([f'<a href="./all-queries.html#all-query-times.{r[0]}.0">The test \'{r[0]}\' is too slow to run as a whole. Investigate whether the create and fill queries can be sped up'])
             else:
                 attrs[6] = ''
 
             if float(r[5]) > allowed_single_run_time * total_runs:
                 slow_average_tests += 1
                 attrs[5] = f'style="background: {color_bad}"'
-                errors_explained.append([f'<a href="./all-queries.html#all-query-times.0">Some query of the test \'{r[0]}\' is too slow to run. See the all queries report'])
+                errors_explained.append([f'<a href="./all-queries.html#all-query-times.{r[0]}.0">Some query of the test \'{r[0]}\' is too slow to run. See the all queries report'])
             else:
                 attrs[5] = ''
 
@@ -659,6 +664,7 @@ elif args.report == 'all-queries':
         attrs[0] = None
         attrs[1] = None
         for r in rows:
+            anchor = f'{currentTableAnchor()}.{r[7]}.{r[8]}'
             if int(r[1]):
                 attrs[6] = f'style="background: {color_bad}"'
             else:
@@ -679,7 +685,7 @@ elif args.report == 'all-queries':
                 attrs[2] = ''
                 attrs[3] = ''
 
-            text += tableRow(r, attrs)
+            text += tableRow(r, attrs, anchor)
 
         text += tableEnd()
         tables.append(text)

docs/en/sql-reference/statements/system.md

@@ -277,8 +277,4 @@ SYSTEM RESTART REPLICA [db.]replicated_merge_tree_family_table_name
 
 Provides possibility to reinitialize Zookeeper sessions state for all `ReplicatedMergeTree` tables, will compare current state with Zookeeper as source of true and add tasks to Zookeeper queue if needed
 
-``` sql
-SYSTEM RESTART QUEUES [db.]replicated_merge_tree_family_table_name
-```
-
 [Original article](https://clickhouse.tech/docs/en/query_language/system/) <!--hide-->

docs/ru/engines/table-engines/integrations/index.md

@@ -3,4 +3,16 @@ toc_folder_title: Integrations
 toc_priority: 30
 ---
 
+# Движки таблиц для интеграции {#table-engines-for-integrations}
 
+Для интеграции с внешними системами ClickHouse предоставляет различные средства, включая движки таблиц. Конфигурирование интеграционных движков осуществляется с помощью запросов `CREATE TABLE` или `ALTER TABLE`, как и для других табличных движков. С точки зрения пользователя, настроенная интеграция выглядит как обычная таблица, но запросы к ней передаются через прокси во внешнюю систему. Этот прозрачный запрос является одним из ключевых преимуществ этого подхода по сравнению с альтернативными методами интеграции, такими как внешние словари или табличные функции, которые требуют использования пользовательских методов запроса при каждом использовании.
+
+Список поддерживаемых интеграций:
+
+-   [ODBC](../../../engines/table-engines/integrations/odbc.md)
+-   [JDBC](../../../engines/table-engines/integrations/jdbc.md)
+-   [MySQL](../../../engines/table-engines/integrations/mysql.md)
+-   [HDFS](../../../engines/table-engines/integrations/hdfs.md)
+-   [Kafka](../../../engines/table-engines/integrations/kafka.md)
+
+[Оригинальная статья](https://clickhouse.tech/docs/ru/engines/table-engines/integrations/) <!--hide-->

docs/ru/engines/table-engines/mergetree-family/replacingmergetree.md

@@ -1,6 +1,6 @@
 # ReplacingMergeTree {#replacingmergetree}
 
-Движок отличается от [MergeTree](mergetree.md#table_engines-mergetree) тем, что выполняет удаление дублирующихся записей с одинаковым значением первичного ключа (точнее, с одинаковым значением [ключа сортировки](mergetree.md)).
+Движок отличается от [MergeTree](mergetree.md#table_engines-mergetree) тем, что выполняет удаление дублирующихся записей с одинаковым значением [ключа сортировки](mergetree.md)).
 
 Дедупликация данных производится лишь во время слияний. Слияние происходят в фоне в неизвестный момент времени, на который вы не можете ориентироваться. Некоторая часть данных может остаться необработанной. Хотя вы можете вызвать внеочередное слияние с помощью запроса `OPTIMIZE`, на это не стоит рассчитывать, так как запрос `OPTIMIZE` приводит к чтению и записи большого объёма данных.
 
@@ -27,7 +27,7 @@ CREATE TABLE [IF NOT EXISTS] [db.]table_name [ON CLUSTER cluster]
 
 -   `ver` — столбец с версией, тип `UInt*`, `Date` или `DateTime`. Необязательный параметр.
 
-        При слиянии, из всех строк с одинаковым значением первичного ключа `ReplacingMergeTree` оставляет только одну:
+        При слиянии, из всех строк с одинаковым значением ключа сортировки `ReplacingMergeTree` оставляет только одну:
 
         - Последнюю в выборке, если `ver` не задан.
         - С максимальной версией, если `ver` задан.
@@ -40,7 +40,7 @@ CREATE TABLE [IF NOT EXISTS] [db.]table_name [ON CLUSTER cluster]
 
 <summary>Устаревший способ создания таблицы</summary>
 
-!!! attention "Attention"
+!!! attention "Внимание"
     Не используйте этот способ в новых проектах и по возможности переведите старые проекты на способ описанный выше.
 
 ``` sql

docs/ru/engines/table-engines/special/index.md

@@ -3,4 +3,14 @@ toc_folder_title: Special
 toc_priority: 31
 ---
 
+# Специальные движки таблиц {#special-table-engines}
 
+Существует три основные категории движков таблиц:
+
+-   [Семейство MergeTree](../../../engines/table-engines/mergetree-family/index.md) для основного использования.
+-   [Семейство Log](../../../engines/table-engines/log-family/index.md) для небольших временных данных.
+-   [Движки таблиц для интеграции](../../../engines/table-engines/integrations/index.md).
+
+Остальные движки таблиц уникальны по своему назначению и еще не сгруппированы в семейства, поэтому они помещены в эту специальную категорию.
+
+[Оригинальная статья](https://clickhouse.tech/docs/ru/engines/table-engines/special/) <!--hide-->

docs/ru/interfaces/formats.md

@@ -7,7 +7,7 @@ ClickHouse может принимать (`INSERT`) и отдавать (`SELECT
 | Формат                                                          | INSERT | SELECT |
 |-----------------------------------------------------------------|--------|--------|
 | [TabSeparated](#tabseparated)                                   | ✔      | ✔      |
-| [TabSeparatedRaw](#tabseparatedraw)                             | ✗      | ✔      |
+| [TabSeparatedRaw](#tabseparatedraw)                             | ✔      | ✔      |
 | [TabSeparatedWithNames](#tabseparatedwithnames)                 | ✔      | ✔      |
 | [TabSeparatedWithNamesAndTypes](#tabseparatedwithnamesandtypes) | ✔      | ✔      |
 | [Template](#format-template)                                    | ✔      | ✔      |
@@ -132,7 +132,7 @@ SELECT * FROM nestedt FORMAT TSV
 ## TabSeparatedRaw {#tabseparatedraw}
 
 Отличается от формата `TabSeparated` тем, что строки выводятся без экранирования.
-Этот формат подходит только для вывода результата выполнения запроса, но не для парсинга (приёма данных для вставки в таблицу).
+Используя этот формат, следите, чтобы в полях не было символов табуляции или разрыва строки.
 
 Этот формат также доступен под именем `TSVRaw`.
 

docs/ru/interfaces/third-party/index.md

@@ -4,4 +4,16 @@ toc_folder_title: "\u041E\u0442 \u0441\u0442\u043E\u0440\u043E\u043D\u043D\u0438
 toc_priority: 24
 ---
 
+# Сторонние интерфейсы {#third-party-interfaces}
 
+Раздел содержит список сторонних интерфейсов для ClickHouse. Это может быть визуальный интерфейс, интерфейс командной строки, либо API:
+
+-   [Client libraries](../../interfaces/third-party/client-libraries.md)
+-   [Integrations](../../interfaces/third-party/integrations.md)
+-   [GUI](../../interfaces/third-party/gui.md)
+-   [Proxies](../../interfaces/third-party/proxy.md)
+
+!!! note "Примечание"
+    С ClickHouse работают также универсальные инструменты, поддерживающие общий API, такие как [ODBC](../../interfaces/odbc.md) или [JDBC](../../interfaces/jdbc.md).
+
+[Оригинальная статья](https://clickhouse.tech/docs/ru/interfaces/third-party/) <!--hide-->

docs/ru/sql-reference/statements/system.md

@@ -235,14 +235,10 @@ SYSTEM SYNC REPLICA [db.]replicated_merge_tree_family_table_name
 Инициализация очереди репликации на основе данных ZooKeeper, происходит так же как при attach table. На короткое время таблица станет недоступной для любых операций.
 
 ``` sql
-SYSTEM RESTART QUEUES [db.]replicated_merge_tree_family_table_name
+SYSTEM RESTART REPLICA [db.]replicated_merge_tree_family_table_name
 ```
 
 ### RESTART REPLICAS {#query_language-system-restart-replicas}
 Реинициализация состояния Zookeeper сессий для всех `ReplicatedMergeTree` таблиц, сравнивает текущее состояние с тем что хранится в Zookeeper как источник правды и добавляет задачи Zookeeper очередь если необходимо
 
-``` sql
-SYSTEM RESTART QUEUES [db.]replicated_merge_tree_family_table_name
-```
-
 [Оригинальная статья](https://clickhouse.tech/docs/ru/query_language/system/) <!--hide-->

docs/zh/sql-reference/statements/system.md

@@ -280,8 +280,4 @@ SYSTEM RESTART REPLICA [db.]replicated_merge_tree_family_table_name
 
 重置所有 `ReplicatedMergeTree`表的ZooKeeper会话状态。该操作会以Zookeeper为参照，对比当前状态，有需要的情况下将任务添加到ZooKeeper队列。
 
-``` sql
-SYSTEM RESTART QUEUES [db.]replicated_merge_tree_family_table_name
-```
-
 [原始文档](https://clickhouse.tech/docs/en/query_language/system/) <!--hide-->

programs/server/.gitignore

@@ -0,0 +1,11 @@
+/access
+/dictionaries_lib
+/flags
+/format_schemas
+/metadata_dropped
+/preprocessed_configs
+/tmp
+/user_files
+
+status
+

programs/server/data/.gitignore

@@ -1,5 +1,3 @@
-*.bin
-*.mrk
 *.txt
 *.dat
 *.idx

programs/server/data/default/.gitignore

@@ -1,2 +0,0 @@
-*.bin
-*.mrk

programs/server/metadata/.gitignore

@@ -0,0 +1 @@
+*.sql

src/Common/BitHelpers.h

@@ -1,6 +1,7 @@
 #pragma once
 
 #include <cstddef>
+#include <cassert>
 #include <type_traits>
 #include <common/defines.h>
 
@@ -11,6 +12,7 @@
   */
 inline unsigned int bitScanReverse(unsigned int x)
 {
+    assert(x != 0);
     return sizeof(unsigned int) * 8 - 1 - __builtin_clz(x);
 }
 

src/Common/PODArray.h

@@ -21,6 +21,15 @@
 
 #include <Common/PODArray_fwd.h>
 
+/** Whether we can use memcpy instead of a loop with assignment to T from U.
+  * It is Ok if types are the same. And if types are integral and of the same size,
+  *  example: char, signed char, unsigned char.
+  * It's not Ok for int and float.
+  * Don't forget to apply std::decay when using this constexpr.
+  */
+template <typename T, typename U>
+constexpr bool memcpy_can_be_used_for_assignment = std::is_same_v<T, U>
+    || (std::is_integral_v<T> && std::is_integral_v<U> && sizeof(T) == sizeof(U));
 
 namespace DB
 {
@@ -313,7 +322,15 @@ public:
         insert(from_begin, from_end);
     }
 
-    PODArray(std::initializer_list<T> il) : PODArray(std::begin(il), std::end(il)) {}
+    PODArray(std::initializer_list<T> il)
+    {
+        this->reserve(std::size(il));
+
+        for (const auto & x : il)
+        {
+            this->push_back(x);
+        }
+    }
 
     PODArray(PODArray && other)
     {
@@ -428,17 +445,21 @@ public:
     void insertSmallAllowReadWriteOverflow15(It1 from_begin, It2 from_end, TAllocatorParams &&... allocator_params)
     {
         static_assert(pad_right_ >= 15);
+        static_assert(sizeof(T) == sizeof(*from_begin));
         insertPrepare(from_begin, from_end, std::forward<TAllocatorParams>(allocator_params)...);
         size_t bytes_to_copy = this->byte_size(from_end - from_begin);
         memcpySmallAllowReadWriteOverflow15(this->c_end, reinterpret_cast<const void *>(&*from_begin), bytes_to_copy);
         this->c_end += bytes_to_copy;
     }
 
+    /// Do not insert into the array a piece of itself. Because with the resize, the iterators on themselves can be invalidated.
     template <typename It1, typename It2>
     void insert(iterator it, It1 from_begin, It2 from_end)
     {
+        static_assert(memcpy_can_be_used_for_assignment<std::decay_t<T>, std::decay_t<decltype(*from_begin)>>);
+
         size_t bytes_to_copy = this->byte_size(from_end - from_begin);
-        size_t bytes_to_move = (end() - it) * sizeof(T);
+        size_t bytes_to_move = this->byte_size(end() - it);
 
         insertPrepare(from_begin, from_end);
 
@@ -446,12 +467,15 @@ public:
             memcpy(this->c_end + bytes_to_copy - bytes_to_move, this->c_end - bytes_to_move, bytes_to_move);
 
         memcpy(this->c_end - bytes_to_move, reinterpret_cast<const void *>(&*from_begin), bytes_to_copy);
+
         this->c_end += bytes_to_copy;
     }
 
     template <typename It1, typename It2>
     void insert_assume_reserved(It1 from_begin, It2 from_end)
     {
+        static_assert(memcpy_can_be_used_for_assignment<std::decay_t<T>, std::decay_t<decltype(*from_begin)>>);
+
         size_t bytes_to_copy = this->byte_size(from_end - from_begin);
         memcpy(this->c_end, reinterpret_cast<const void *>(&*from_begin), bytes_to_copy);
         this->c_end += bytes_to_copy;
@@ -584,12 +608,15 @@ public:
     template <typename It1, typename It2, typename... TAllocatorParams>
     void assign(It1 from_begin, It2 from_end, TAllocatorParams &&... allocator_params)
     {
+        static_assert(memcpy_can_be_used_for_assignment<std::decay_t<T>, std::decay_t<decltype(*from_begin)>>);
+
         size_t required_capacity = from_end - from_begin;
         if (required_capacity > this->capacity())
             this->reserve(roundUpToPowerOfTwoOrZero(required_capacity), std::forward<TAllocatorParams>(allocator_params)...);
 
         size_t bytes_to_copy = this->byte_size(required_capacity);
         memcpy(this->c_start, reinterpret_cast<const void *>(&*from_begin), bytes_to_copy);
+
         this->c_end = this->c_start + bytes_to_copy;
     }
 

src/Functions/FunctionsComparison.h

@@ -13,6 +13,7 @@
 
 #include <DataTypes/DataTypesNumber.h>
 #include <DataTypes/DataTypeNullable.h>
+#include <DataTypes/DataTypeNothing.h>
 #include <DataTypes/DataTypeDateTime.h>
 #include <DataTypes/DataTypeDateTime64.h>
 #include <DataTypes/DataTypeDate.h>
@@ -884,11 +885,18 @@ private:
         if (tuple_size != typeid_cast<const DataTypeTuple &>(*c1.type).getElements().size())
             throw Exception("Cannot compare tuples of different sizes.", ErrorCodes::BAD_ARGUMENTS);
 
+        auto & res = block.getByPosition(result);
+        if (res.type->onlyNull())
+        {
+            res.column = res.type->createColumnConstWithDefaultValue(input_rows_count);
+            return;
+        }
+
         ColumnsWithTypeAndName x(tuple_size);
         ColumnsWithTypeAndName y(tuple_size);
 
-        auto x_const = checkAndGetColumnConst<ColumnTuple>(c0.column.get());
+        const auto * x_const = checkAndGetColumnConst<ColumnTuple>(c0.column.get());
-        auto y_const = checkAndGetColumnConst<ColumnTuple>(c1.column.get());
+        const auto * y_const = checkAndGetColumnConst<ColumnTuple>(c1.column.get());
 
         Columns x_columns;
         Columns y_columns;
@@ -1135,17 +1143,22 @@ public:
                 FunctionComparison<Op, Name>::create(context)));
 
             bool has_nullable = false;
+            bool has_null = false;
 
             size_t size = left_tuple->getElements().size();
             for (size_t i = 0; i < size; ++i)
             {
                 ColumnsWithTypeAndName args = {{nullptr, left_tuple->getElements()[i], ""},
                                                {nullptr, right_tuple->getElements()[i], ""}};
-                has_nullable = has_nullable || adaptor.build(args)->getReturnType()->isNullable();
+                auto element_type = adaptor.build(args)->getReturnType();
+                has_nullable = has_nullable || element_type->isNullable();
+                has_null = has_null || element_type->onlyNull();
             }
 
             /// If any element comparison is nullable, return type will also be nullable.
             /// We useDefaultImplementationForNulls, but it doesn't work for tuples.
+            if (has_null)
+                return std::make_shared<DataTypeNullable>(std::make_shared<DataTypeNothing>());
             if (has_nullable)
                 return std::make_shared<DataTypeNullable>(std::make_shared<DataTypeUInt8>());
         }

src/IO/parseDateTimeBestEffort.cpp

@@ -104,6 +104,7 @@ ReturnType parseDateTimeBestEffortImpl(
             return false;
     };
 
+    res = 0;
     UInt16 year = 0;
     UInt8 month = 0;
     UInt8 day_of_month = 0;

src/IO/readDecimalText.h

@@ -1,9 +1,29 @@
 #pragma once
 
+#include <limits>
 #include <IO/ReadHelpers.h>
 #include <Common/intExp.h>
 
 
+/// This is only needed for non-official, "unbundled" build.
+/// https://stackoverflow.com/questions/41198673/uint128-t-not-working-with-clang-and-libstdc
+#if !defined(_LIBCPP_LIMITS) && !defined(__GLIBCXX_BITSIZE_INT_N_0) && defined(__SIZEOF_INT128__)
+namespace std
+{
+    template <>
+    struct numeric_limits<__int128_t>
+    {
+        static constexpr bool is_specialized = true;
+        static constexpr bool is_signed = true;
+        static constexpr bool is_integer = true;
+        static constexpr int radix = 2;
+        static constexpr int digits = 127;
+        static constexpr int digits10 = 38;
+    };
+}
+#endif
+
+
 namespace DB
 {
 
@@ -160,12 +180,24 @@ inline void readDecimalText(ReadBuffer & buf, T & x, uint32_t precision, uint32_
 
     if (static_cast<int32_t>(scale) + exponent < 0)
     {
-        /// Too many digits after point. Just cut off excessive digits.
+        auto divisor_exp = -exponent - static_cast<int32_t>(scale);
-        auto divisor = intExp10OfSize<T>(-exponent - static_cast<int32_t>(scale));
+
-        assert(divisor > 0);    /// This is for Clang Static Analyzer. It is not smart enough to infer it automatically.
+        if (divisor_exp >= std::numeric_limits<typename T::NativeType>::digits10)
-        x.value /= divisor;
+        {
-        scale = 0;
+            /// Too big negative exponent
-        return;
+            x.value = 0;
+            scale = 0;
+            return;
+        }
+        else
+        {
+            /// Too many digits after point. Just cut off excessive digits.
+            auto divisor = intExp10OfSize<T>(divisor_exp);
+            assert(divisor > 0);    /// This is for Clang Static Analyzer. It is not smart enough to infer it automatically.
+            x.value /= divisor;
+            scale = 0;
+            return;
+        }
     }
 
     scale += exponent;

src/Processors/Formats/Impl/ArrowBlockOutputFormat.cpp

@@ -43,13 +43,17 @@ void ArrowBlockOutputFormat::consume(Chunk chunk)
 
 void ArrowBlockOutputFormat::finalize()
 {
-    if (writer)
+    if (!writer)
     {
-        auto status = writer->Close();
+        const Block & header = getPort(PortKind::Main).getHeader();
-        if (!status.ok())
+
-            throw Exception(ErrorCodes::UNKNOWN_EXCEPTION,
+        consume(Chunk(header.getColumns(), 0));
-                "Error while closing a table: {}", status.ToString());
     }
+
+    auto status = writer->Close();
+    if (!status.ok())
+        throw Exception(ErrorCodes::UNKNOWN_EXCEPTION,
+            "Error while closing a table: {}", status.ToString());
 }
 
 void ArrowBlockOutputFormat::prepareWriter(const std::shared_ptr<arrow::Schema> & schema)

src/Processors/Formats/Impl/ParquetBlockOutputFormat.cpp

@@ -66,12 +66,16 @@ void ParquetBlockOutputFormat::consume(Chunk chunk)
 
 void ParquetBlockOutputFormat::finalize()
 {
-    if (file_writer)
+    if (!file_writer)
     {
-        auto status = file_writer->Close();
+        const Block & header = getPort(PortKind::Main).getHeader();
-        if (!status.ok())
+
-            throw Exception{"Error while closing a table: " + status.ToString(), ErrorCodes::UNKNOWN_EXCEPTION};
+        consume(Chunk(header.getColumns(), 0));
     }
+
+    auto status = file_writer->Close();
+    if (!status.ok())
+        throw Exception{"Error while closing a table: " + status.ToString(), ErrorCodes::UNKNOWN_EXCEPTION};
 }
 
 void registerOutputFormatProcessorParquet(FormatFactory & factory)

src/Processors/Transforms/FinishSortingTransform.cpp

@@ -25,23 +25,16 @@ FinishSortingTransform::FinishSortingTransform(
     const SortDescription & description_to_sort_,
     size_t max_merged_block_size_, UInt64 limit_)
     : SortingTransform(header, description_to_sort_, max_merged_block_size_, limit_)
-    , description_sorted(description_sorted_)
 {
-    const auto & sample = inputs.front().getHeader();
+    /// Check for sanity non-modified descriptions
-
+    if (!isPrefix(description_sorted_, description_to_sort_))
-    /// Replace column names to column position in description_sorted.
+        throw Exception("Can't finish sorting. SortDescription of already sorted stream is not prefix of "
-    for (auto & column_description : description_sorted)
-    {
-        if (!column_description.column_name.empty())
-        {
-            column_description.column_number = sample.getPositionByName(column_description.column_name);
-            column_description.column_name.clear();
-        }
-    }
-
-    if (!isPrefix(description_sorted, description))
-        throw Exception("Can`t finish sorting. SortDescription of already sorted stream is not prefix of "
             "SortDescription needed to sort", ErrorCodes::LOGICAL_ERROR);
+
+    /// The target description is modified in SortingTransform constructor.
+    /// To avoid doing the same actions with description_sorted just copy it from prefix of target description.
+    size_t prefix_size = description_sorted_.size();
+    description_sorted.assign(description.begin(), description.begin() + prefix_size);
 }
 
 static bool less(const Columns & lhs, const Columns & rhs, size_t i, size_t j, const SortDescription & descr)

src/Storages/MergeTree/MergeTreeDataSelectExecutor.cpp

@@ -34,6 +34,7 @@ namespace std
         static constexpr bool is_integer = true;
         static constexpr int radix = 2;
         static constexpr int digits = 128;
+        static constexpr int digits10 = 38;
         static constexpr __uint128_t min () { return 0; } // used in boost 1.65.1+
         static constexpr __uint128_t max () { return __uint128_t(0) - 1; } // used in boost 1.68.0+
     };

src/Storages/System/StorageSystemContributors.generated.cpp

@@ -52,6 +52,7 @@ const char * auto_contributors[] {
     "Aliaksandr Shylau",
     "Amos Bird",
     "Amy Krishnevsky",
+    "AnaUvarova",
     "Anastasiya Rodigina",
     "Anastasiya Tsarkova",
     "Anatoly Pugachev",
@@ -72,8 +73,10 @@ const char * auto_contributors[] {
     "Andrey Skobtsov",
     "Andrey Urusov",
     "Andy Yang",
+    "Anna",
     "Anton Ivashkin",
     "Anton Kobzev",
+    "Anton Kvasha",
     "Anton Okhitin",
     "Anton Okulov",
     "Anton Patsev",
@@ -90,6 +93,7 @@ const char * auto_contributors[] {
     "Artem Zuikov",
     "Artemeey",
     "Artemkin Pavel",
+    "Arthur Petukhovsky",
     "Arthur Tokarchuk",
     "Artur Beglaryan",
     "AsiaKorushkina",
@@ -104,6 +108,7 @@ const char * auto_contributors[] {
     "BayoNet",
     "Bharat Nallan",
     "Big Elephant",
+    "Bill",
     "BlahGeek",
     "Bogdan",
     "Bogdan Voronin",
@@ -130,6 +135,7 @@ const char * auto_contributors[] {
     "Darío",
     "Denis Burlaka",
     "Denis Glazachev",
+    "Denis Krivak",
     "Denis Zhuravlev",
     "Derek Perkins",
     "Ding Xiang Fei",
@@ -152,6 +158,7 @@ const char * auto_contributors[] {
     "Eldar Zaitov",
     "Elena Baskakova",
     "Elghazal Ahmed",
+    "Elizaveta Mironyuk",
     "Emmanuel Donin de Rosière",
     "Eric",
     "Ernest Poletaev",
@@ -259,6 +266,7 @@ const char * auto_contributors[] {
     "Marek Vavrusa",
     "Marek Vavruša",
     "Marek Vavruša",
+    "Mark Papadakis",
     "Maroun Maroun",
     "Marsel Arduanov",
     "Marti Raudsepp",
@@ -307,6 +315,7 @@ const char * auto_contributors[] {
     "Mohammad Hossein Sekhavat",
     "MovElb",
     "Murat Kabilov",
+    "MyroTk",
     "NIKITA MIKHAILOV",
     "Narek Galstyan",
     "NeZeD [Mac Pro]",
@@ -346,6 +355,7 @@ const char * auto_contributors[] {
     "Pavel Yakunin",
     "Pavlo Bashynskiy",
     "Pawel Rog",
+    "Peng Jian",
     "Persiyanov Dmitriy Andreevich",
     "Pervakov Grigorii",
     "Pervakov Grigory",
@@ -359,6 +369,7 @@ const char * auto_contributors[] {
     "Reilee",
     "Reto Kromer",
     "Ri",
+    "Roman Bug",
     "Roman Lipovsky",
     "Roman Nikolaev",
     "Roman Nozdrin",
@@ -466,6 +477,7 @@ const char * auto_contributors[] {
     "Yurii Vlasenko",
     "Yuriy",
     "Yuriy Baranov",
+    "Yuriy Chernyshov",
     "Yury Karpovich",
     "Yury Stankevich",
     "Zhichang Yu",
@@ -483,6 +495,7 @@ const char * auto_contributors[] {
     "alex.lvxin",
     "alexander kozhikhov",
     "alexey-milovidov",
+    "amoschen",
     "amudong",
     "andrei-karpliuk",
     "andrewsg",
@@ -556,6 +569,7 @@ const char * auto_contributors[] {
     "imgbot[bot]",
     "ivan-kush",
     "ivanzhukov",
+    "jakalletti",
     "javartisan",
     "javi",
     "javi santana",
@@ -610,6 +624,7 @@ const char * auto_contributors[] {
     "objatie_groba",
     "ogorbacheva",
     "olegkv",
+    "olgarev",
     "orantius",
     "palasonicq",
     "peshkurov",
@@ -621,6 +636,7 @@ const char * auto_contributors[] {
     "qianlixiang",
     "quid",
     "rainbowsysu",
+    "ritaank",
     "robot-clickhouse",
     "robot-metrika-test",
     "root",
@@ -652,6 +668,7 @@ const char * auto_contributors[] {
     "vinity",
     "vitstn",
     "vivarum",
+    "vladimir golovchenko",
     "vxider",
     "vzakaznikov",
     "wangchao",
@@ -679,6 +696,7 @@ const char * auto_contributors[] {
     "张风啸",
     "极客青年",
     "谢磊",
+    "贾顺名(Jarvis)",
     "黄朝晖",
     "黄璞",
     "박현우",

tests/queries/0_stateless/00250_tuple_comparison.reference

@@ -13,3 +13,4 @@
 1	0	0	0	1	1
 0	1	1	0	1	0
 0	1	0	1	0	1
+\N	\N	\N	\N	\N

tests/queries/0_stateless/00250_tuple_comparison.sql

@@ -103,3 +103,9 @@ SELECT
     tuple(2) >  tuple(1),
     tuple(2) <= tuple(1),
     tuple(2) >= tuple(1);
+SELECT
+    tuple(NULL) < tuple(1),
+    tuple(NULL) = tuple(1),
+    tuple(NULL) <= tuple(1),
+    tuple(1, NULL) = tuple(2, 1),
+    tuple(1, NULL) < tuple(2, 1);

tests/queries/0_stateless/01425_decimal_parse_big_negative_exponent.reference

@@ -0,0 +1,6 @@
+1E-9	0
+1E-8	0
+1E-7	0
+1e-7	0
+1E-9	0.000000001
+1E-10	0.000000000

tests/queries/0_stateless/01425_decimal_parse_big_negative_exponent.sql

@@ -0,0 +1,10 @@
+SELECT '-1E9-1E9-1E9-1E9' AS x, toDecimal32(x, 0); -- { serverError 6 }
+SELECT '-1E9' AS x, toDecimal32(x, 0); -- { serverError 69 }
+SELECT '1E-9' AS x, toDecimal32(x, 0);
+SELECT '1E-8' AS x, toDecimal32(x, 0);
+SELECT '1E-7' AS x, toDecimal32(x, 0);
+SELECT '1e-7' AS x, toDecimal32(x, 0);
+SELECT '1E-9' AS x, toDecimal32(x, 9);
+SELECT '1E-9' AS x, toDecimal32(x, 10); -- { serverError 69 }
+SELECT '1E-10' AS x, toDecimal32(x, 10); -- { serverError 69 }
+SELECT '1E-10' AS x, toDecimal32(x, 9);

tests/queries/0_stateless/01429_empty_arrow_and_parquet.reference

@@ -0,0 +1,6 @@
+1
+1
+2
+2
+3
+3

tests/queries/0_stateless/01429_empty_arrow_and_parquet.sh

@@ -0,0 +1,23 @@
+#!/usr/bin/env bash
+
+set -e
+
+CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
+. "$CUR_DIR"/../shell_config.sh
+
+
+${CLICKHOUSE_CLIENT} --query="DROP TABLE IF EXISTS test_empty_data"
+${CLICKHOUSE_CLIENT} --query="CREATE TABLE test_empty_data (x Int8) ENGINE = Memory"
+
+(echo "INSERT INTO test_empty_data FORMAT Arrow" && ${CLICKHOUSE_CLIENT} --query="SELECT 1 AS x FORMAT Arrow") | ${CLICKHOUSE_CLIENT}
+${CLICKHOUSE_CLIENT} --query="SELECT count() FROM test_empty_data"
+(echo "INSERT INTO test_empty_data FORMAT Arrow" && ${CLICKHOUSE_CLIENT} --query="SELECT 1 AS x LIMIT 0 FORMAT Arrow") | ${CLICKHOUSE_CLIENT}
+${CLICKHOUSE_CLIENT} --query="SELECT count() FROM test_empty_data"
+(echo "INSERT INTO test_empty_data FORMAT ArrowStream" && ${CLICKHOUSE_CLIENT} --query="SELECT 1 AS x FORMAT ArrowStream") | ${CLICKHOUSE_CLIENT}
+${CLICKHOUSE_CLIENT} --query="SELECT count() FROM test_empty_data"
+(echo "INSERT INTO test_empty_data FORMAT ArrowStream" && ${CLICKHOUSE_CLIENT} --query="SELECT 1 AS x LIMIT 0 FORMAT ArrowStream") | ${CLICKHOUSE_CLIENT}
+${CLICKHOUSE_CLIENT} --query="SELECT count() FROM test_empty_data"
+(echo "INSERT INTO test_empty_data FORMAT Parquet" && ${CLICKHOUSE_CLIENT} --query="SELECT 1 AS x FORMAT Parquet") | ${CLICKHOUSE_CLIENT}
+${CLICKHOUSE_CLIENT} --query="SELECT count() FROM test_empty_data"
+(echo "INSERT INTO test_empty_data FORMAT Parquet" && ${CLICKHOUSE_CLIENT} --query="SELECT 1 AS x LIMIT 0 FORMAT Parquet") | ${CLICKHOUSE_CLIENT}
+${CLICKHOUSE_CLIENT} --query="SELECT count() FROM test_empty_data"

tests/queries/0_stateless/01431_finish_sorting_with_consts.reference

@@ -0,0 +1,3 @@
+1	2020-05-05 01:00:00	0
+1	2020-05-05 01:00:00	1
+1	2020-05-05 01:00:00	2

tests/queries/0_stateless/01431_finish_sorting_with_consts.sql

@@ -0,0 +1,7 @@
+DROP TABLE IF EXISTS pk_func;
+
+CREATE TABLE pk_func (`d` DateTime, `ui` UInt32 ) ENGINE = MergeTree ORDER BY toDate(d);
+INSERT INTO pk_func SELECT '2020-05-05 01:00:00', number FROM numbers(1000);
+SELECT 1, * FROM pk_func ORDER BY toDate(d) ASC, ui ASC LIMIT 3;
+
+DROP TABLE IF EXISTS pk_func;

tests/queries/0_stateless/01432_parse_date_time_best_effort_timestamp.reference

@@ -0,0 +1,3 @@
+2020-08-07 01:29:00
+1973-03-03 12:46:40
+2020-08-07 00:00:00

tests/queries/0_stateless/01432_parse_date_time_best_effort_timestamp.sql

@@ -0,0 +1,3 @@
+SELECT parseDateTimeBestEffort('1596752940', 'Europe/Moscow');
+SELECT parseDateTimeBestEffort('100000000', 'Europe/Moscow');
+SELECT parseDateTimeBestEffort('20200807', 'Europe/Moscow');

tests/queries/0_stateless/01433_hex_float.reference

@@ -0,0 +1,2 @@
+0.123
+1

tests/queries/0_stateless/01433_hex_float.sql

@@ -0,0 +1,4 @@
+CREATE TEMPORARY TABLE t (x Float64);
+INSERT INTO t VALUES (0x1.f7ced916872b0p-4);
+SELECT * FROM t;
+SELECT x = 0x1.f7ced916872b0p-4 FROM t;

website/js/docs.js

@@ -49,9 +49,7 @@ $(document).ready(function () {
     $('#sidebar .nav-link.active').parents('.collapse').each(function() {
         var current = $(this);
         if (current.attr('id') !== 'sidebar') {
-            current.css('transition-duration', '0s');
             current.collapse('show');
-            current.css('transition-duration', '0.4s');
         }
     });
     $(window).resize(onResize);