Merge branch 'master' into yandex-to-clickhouse-in-configs

src/Common/Config/YAMLParser.cpp

@@ -156,7 +156,7 @@ Poco::AutoPtr<Poco::XML::Document> YAMLParser::parse(const String& path)
         throw Exception(ErrorCodes::CANNOT_OPEN_FILE, "Unable to open YAML configuration file {}", path);
     }
     Poco::AutoPtr<Poco::XML::Document> xml = new Document;
-    Poco::AutoPtr<Poco::XML::Element> root_node = xml->createElement("yandex");
+    Poco::AutoPtr<Poco::XML::Element> root_node = xml->createElement("clickhouse");
     xml->appendChild(root_node);
     try
     {

src/Common/FileChecker.cpp

@@ -111,7 +111,7 @@ void FileChecker::save() const
         std::unique_ptr<WriteBuffer> out = disk->writeFile(tmp_files_info_path);
 
         /// So complex JSON structure - for compatibility with the old format.
-        writeCString("{\"yandex\":{", *out);
+        writeCString("{\"clickhouse\":{", *out);
 
         auto settings = FormatSettings();
         for (auto it = map.begin(); it != map.end(); ++it)
@@ -153,7 +153,7 @@ void FileChecker::load()
     }
     JSON json(out.str());
 
-    JSON files = json["yandex"];
+    JSON files = json.has("clickhouse") ? json["clickhouse"] : json["yandex"];
     for (const JSON file : files) // NOLINT
         map[unescapeForFileName(file.getName())] = file.getValue()["size"].toUInt();
 }

src/Interpreters/BloomFilter.cpp

@@ -1,9 +1,11 @@
 #include <Interpreters/BloomFilter.h>
 #include <city.h>
 #include <Columns/ColumnArray.h>
+#include <Columns/ColumnMap.h>
 #include <Columns/ColumnNullable.h>
 #include <Columns/ColumnLowCardinality.h>
 #include <DataTypes/DataTypeArray.h>
+#include <DataTypes/DataTypeMap.h>
 #include <DataTypes/DataTypeNullable.h>
 #include <DataTypes/DataTypeLowCardinality.h>
 
@@ -124,6 +126,14 @@ DataTypePtr BloomFilter::getPrimitiveType(const DataTypePtr & data_type)
             throw Exception("Unexpected type " + data_type->getName() + " of bloom filter index.", ErrorCodes::BAD_ARGUMENTS);
     }
 
+    if (const auto * map_type = typeid_cast<const DataTypeMap *>(data_type.get()))
+    {
+        if (!typeid_cast<const DataTypeMap *>(map_type->getKeyType().get()))
+            return getPrimitiveType(map_type->getKeyType());
+        else
+            throw Exception("Unexpected key type " + data_type->getName() + " of bloom filter index for map.", ErrorCodes::BAD_ARGUMENTS);
+    }
+
     if (const auto * nullable_type = typeid_cast<const DataTypeNullable *>(data_type.get()))
         return getPrimitiveType(nullable_type->getNestedType());
 
@@ -138,6 +148,9 @@ ColumnPtr BloomFilter::getPrimitiveColumn(const ColumnPtr & column)
     if (const auto * array_col = typeid_cast<const ColumnArray *>(column.get()))
         return getPrimitiveColumn(array_col->getDataPtr());
 
+    if (const auto * map_col = typeid_cast<const ColumnMap *>(column.get()))
+        return getPrimitiveColumn(map_col->getNestedData().getColumnPtr(0));
+
     if (const auto * nullable_col = typeid_cast<const ColumnNullable *>(column.get()))
         return getPrimitiveColumn(nullable_col->getNestedColumnPtr());
 

src/Interpreters/BloomFilterHash.h

@@ -5,6 +5,7 @@
 #include <Columns/IColumn.h>
 #include <Columns/ColumnArray.h>
 #include <Columns/ColumnConst.h>
+#include <Columns/ColumnMap.h>
 #include <Columns/ColumnNullable.h>
 #include <Columns/ColumnsNumber.h>
 #include <Columns/ColumnString.h>
@@ -123,6 +124,26 @@ struct BloomFilterHash
             }
         }
 
+        if (which.isMap())
+        {
+            const auto * map_col = typeid_cast<const ColumnMap *>(column.get());
+            const auto & keys_data = map_col->getNestedData().getColumn(0);
+            if (checkAndGetColumn<ColumnNullable>(keys_data))
+                throw Exception("Unexpected key type " + data_type->getName() + " of bloom filter index.", ErrorCodes::BAD_ARGUMENTS);
+
+            const auto & offsets = map_col->getNestedColumn().getOffsets();
+            limit = offsets[pos + limit - 1] - offsets[pos - 1];  /// PaddedPODArray allows access on index -1.
+            pos = offsets[pos - 1];
+
+            if (limit == 0)
+            {
+                auto index_column = ColumnUInt64::create(1);
+                ColumnUInt64::Container & index_column_vec = index_column->getData();
+                index_column_vec[0] = 0;
+                return index_column;
+            }
+        }
+
         const ColumnPtr actual_col = BloomFilter::getPrimitiveColumn(column);
         const DataTypePtr actual_type = BloomFilter::getPrimitiveType(data_type);
 

src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp

@@ -1,6 +1,7 @@
 #include <Common/HashTable/ClearableHashMap.h>
 #include <Common/FieldVisitorsAccurateComparison.h>
 #include <DataTypes/DataTypeArray.h>
+#include <DataTypes/DataTypeMap.h>
 #include <DataTypes/DataTypeTuple.h>
 #include <Columns/ColumnConst.h>
 #include <Columns/ColumnTuple.h>
@@ -188,7 +189,6 @@ bool MergeTreeIndexConditionBloomFilter::mayBeTrueOnGranule(const MergeTreeIndex
                 const auto & filter = filters[query_index_hash.first];
                 const ColumnPtr & hash_column = query_index_hash.second;
 
-
                 match_rows = maybeTrueOnBloomFilter(&*hash_column,
                                                     filter,
                                                     hash_functions,
@@ -509,6 +509,37 @@ bool MergeTreeIndexConditionBloomFilter::traverseASTEquals(
 
             return match_with_subtype;
         }
+
+        if (function->name == "arrayElement")
+        {
+            const auto & col_name = assert_cast<ASTIdentifier *>(function->arguments.get()->children[0].get())->name();
+
+            if (header.has(col_name))
+            {
+                size_t position = header.getPositionByName(col_name);
+                const DataTypePtr & index_type = header.getByPosition(position).type;
+                const auto * map_type = typeid_cast<const DataTypeMap *>(index_type.get());
+                if (map_type)
+                {
+                    out.function = function_name == "equals" ? RPNElement::FUNCTION_EQUALS : RPNElement::FUNCTION_NOT_EQUALS;
+
+                    auto & argument = function->arguments.get()->children[1];
+
+                    if (const auto * literal = argument->as<ASTLiteral>())
+                    {
+                        auto element_key = literal->value;
+                        const DataTypePtr actual_type = BloomFilter::getPrimitiveType(index_type);
+                        out.predicate.emplace_back(std::make_pair(position, BloomFilterHash::hashWithField(actual_type.get(), element_key)));
+                    }
+                    else
+                    {
+                        return false;
+                    }
+
+                    return true;
+                }
+            }
+        }
     }
 
     return false;