#include #include #include #include #include #include #include namespace { using namespace DB; Field executeFunctionOnField( const Field & field, const std::string & name, const ExpressionActionsPtr & sharding_expr, const DataTypePtr & type, const std::string & sharding_key_column_name) { ColumnWithTypeAndName column; column.column = type->createColumnConst(1, field); column.name = name; column.type = type; Block block{column}; size_t num_rows = 1; sharding_expr->execute(block, num_rows); ColumnWithTypeAndName & ret = block.getByName(sharding_key_column_name); return (*ret.column)[0]; } /// @param sharding_column_value - one of values from IN /// @param sharding_column_name - name of that column /// @return true if shard may contain such value (or it is unknown), otherwise false. bool shardContains( Field sharding_column_value, const std::string & sharding_column_name, const OptimizeShardingKeyRewriteInMatcher::Data & data) { /// Implicit conversion. sharding_column_value = convertFieldToType(sharding_column_value, *data.sharding_key_type); /// NULL is not allowed in sharding key, /// so it should be safe to assume that shard cannot contain it. if (sharding_column_value.isNull()) return false; Field sharding_value = executeFunctionOnField( sharding_column_value, sharding_column_name, data.sharding_key_expr, data.sharding_key_type, data.sharding_key_column_name); /// The value from IN can be non-numeric, /// but in this case it should be convertible to numeric type, let's try. /// /// NOTE: that conversion should not be done for signed types, /// since it uses accurate cast, that will return Null, /// but we need static_cast<> (as createBlockSelector()). if (!isInt64OrUInt64FieldType(sharding_value.getType())) sharding_value = convertFieldToType(sharding_value, DataTypeUInt64()); /// In case of conversion is not possible (NULL), shard cannot contain the value anyway. if (sharding_value.isNull()) return false; UInt64 value = sharding_value.get(); const auto shard_num = data.slots[value % data.slots.size()] + 1; return data.shard_info.shard_num == shard_num; } } namespace DB { bool OptimizeShardingKeyRewriteInMatcher::needChildVisit(ASTPtr & /*node*/, const ASTPtr & /*child*/) { return true; } void OptimizeShardingKeyRewriteInMatcher::visit(ASTPtr & node, Data & data) { if (auto * function = node->as()) visit(*function, data); } void OptimizeShardingKeyRewriteInMatcher::visit(ASTFunction & function, Data & data) { if (function.name != "in") return; auto * left = function.arguments->children.front().get(); auto * right = function.arguments->children.back().get(); auto * identifier = left->as(); if (!identifier) return; if (!data.sharding_key_expr->getRequiredColumnsWithTypes().contains(identifier->name())) return; /// NOTE: that we should not take care about empty tuple, /// since after optimize_skip_unused_shards, /// at least one element should match each shard. if (auto * tuple_func = right->as(); tuple_func && tuple_func->name == "tuple") { auto * tuple_elements = tuple_func->children.front()->as(); std::erase_if(tuple_elements->children, [&](auto & child) { auto * literal = child->template as(); return literal && !shardContains(literal->value, identifier->name(), data); }); } else if (auto * tuple_literal = right->as(); tuple_literal && tuple_literal->value.getType() == Field::Types::Tuple) { auto & tuple = tuple_literal->value.get(); std::erase_if(tuple, [&](auto & child) { return !shardContains(child, identifier->name(), data); }); } } }