Manual cherry-pick of #6780

ClickHouse · Oct 28, 2019 · ba8c698 · ba8c698
1 parent ddb7e31
commit ba8c698
Show file tree

Hide file tree

Showing 4 changed files with 60 additions and 25 deletions.
diff --git a/dbms/src/Storages/MergeTree/MergedColumnOnlyOutputStream.cpp b/dbms/src/Storages/MergeTree/MergedColumnOnlyOutputStream.cpp
@@ -19,32 +19,24 @@ MergedColumnOnlyOutputStream::MergedColumnOnlyOutputStream(
     header(header_), part_path(part_path_), sync(sync_), skip_offsets(skip_offsets_),
     already_written_offset_columns(already_written_offset_columns_)
 {
-}
+    serialization_states.reserve(header.columns());
+    WrittenOffsetColumns tmp_offset_columns;
+    IDataType::SerializeBinaryBulkSettings settings;
 
-void MergedColumnOnlyOutputStream::write(const Block & block)
-{
-    if (!initialized)
+    for (const auto & column_name : header.getNames())
     {
-        column_streams.clear();
-        serialization_states.clear();
-        serialization_states.reserve(block.columns());
-        WrittenOffsetColumns tmp_offset_columns;
-        IDataType::SerializeBinaryBulkSettings settings;
-
-        for (size_t i = 0; i < block.columns(); ++i)
-        {
-            const auto & col = block.safeGetByPosition(i);
-
-            const auto columns = storage.getColumns();
-            addStreams(part_path, col.name, *col.type, columns.getCodecOrDefault(col.name, codec), 0, skip_offsets);
-            serialization_states.emplace_back(nullptr);
-            settings.getter = createStreamGetter(col.name, tmp_offset_columns, false);
-            col.type->serializeBinaryBulkStatePrefix(settings, serialization_states.back());
-        }
-
-        initialized = true;
+        const auto & col = header.getByName(column_name);
+
+        const auto columns = storage.getColumns();
+        addStreams(part_path, col.name, *col.type, columns.getCodecOrDefault(col.name, codec), 0, skip_offsets);
+        serialization_states.emplace_back(nullptr);
+        settings.getter = createStreamGetter(col.name, tmp_offset_columns, false);
+        col.type->serializeBinaryBulkStatePrefix(settings, serialization_states.back());
     }
+}
 
+void MergedColumnOnlyOutputStream::write(const Block & block)
+{
     size_t rows = block.rows();
     if (!rows)
         return;
@@ -83,7 +75,7 @@ MergeTreeData::DataPart::Checksums MergedColumnOnlyOutputStream::writeSuffixAndG
         column.type->serializeBinaryBulkStateSuffix(serialize_settings, serialization_states[i]);
 
 
-        if (with_final_mark)
+        if (with_final_mark && (index_offset != 0 || current_mark != 0))
             writeFinalMark(column.name, column.type, offset_columns, skip_offsets, serialize_settings.path);
     }
 
@@ -100,7 +92,6 @@ MergeTreeData::DataPart::Checksums MergedColumnOnlyOutputStream::writeSuffixAndG
 
     column_streams.clear();
     serialization_states.clear();
-    initialized = false;
 
     return checksums;
 }

diff --git a/dbms/src/Storages/MergeTree/MergedColumnOnlyOutputStream.h b/dbms/src/Storages/MergeTree/MergedColumnOnlyOutputStream.h
@@ -29,7 +29,6 @@ class MergedColumnOnlyOutputStream final : public IMergedBlockOutputStream
     SerializationStates serialization_states;
     String part_path;
 
-    bool initialized = false;
     bool sync;
     bool skip_offsets;
 

diff --git a/dbms/tests/queries/0_stateless/01006_simpod_empty_part_single_column_write.reference b/dbms/tests/queries/0_stateless/01006_simpod_empty_part_single_column_write.reference
@@ -0,0 +1,2 @@
+1
+1
diff --git a/dbms/tests/queries/0_stateless/01006_simpod_empty_part_single_column_write.sh b/dbms/tests/queries/0_stateless/01006_simpod_empty_part_single_column_write.sh
@@ -0,0 +1,43 @@
+#!/usr/bin/env bash
+
+CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
+. $CURDIR/../shell_config.sh
+
+. $CURDIR/mergetree_mutations.lib
+
+
+${CLICKHOUSE_CLIENT} --query="DROP TABLE IF EXISTS table_with_empty_part"
+
+${CLICKHOUSE_CLIENT} --query="CREATE TABLE table_with_empty_part
+(
+    id UInt64,
+    value UInt64
+)
+ENGINE = MergeTree()
+ORDER BY id
+PARTITION BY id
+SETTINGS vertical_merge_algorithm_min_rows_to_activate=0, vertical_merge_algorithm_min_columns_to_activate=0
+"
+
+
+${CLICKHOUSE_CLIENT} --query="INSERT INTO table_with_empty_part VALUES (1, 1)"
+
+${CLICKHOUSE_CLIENT} --query="INSERT INTO table_with_empty_part VALUES (2, 2)"
+
+${CLICKHOUSE_CLIENT} --query="ALTER TABLE table_with_empty_part DELETE WHERE id % 2 == 0"
+
+sleep 0.5
+
+mutation_id=`${CLICKHOUSE_CLIENT} --query="SELECT max(mutation_id) FROM system.mutations WHERE table='table_with_empty_part'"`
+
+wait_for_mutation "table_with_empty_part" "$mutation_id"
+
+${CLICKHOUSE_CLIENT} --query="SELECT COUNT(DISTINCT value) FROM table_with_empty_part"
+
+${CLICKHOUSE_CLIENT} --query="ALTER TABLE table_with_empty_part MODIFY COLUMN value Nullable(UInt64)"
+
+${CLICKHOUSE_CLIENT} --query="SELECT COUNT(distinct value) FROM table_with_empty_part"
+
+${CLICKHOUSE_CLIENT} --query="OPTIMIZE TABLE table_with_empty_part FINAL"
+
+${CLICKHOUSE_CLIENT} --query="DROP TABLE IF EXISTS table_with_empty_part"