Release notes of AnalyticDB for MySQL - AnalyticDB - Alibaba Cloud Documentation Center

This topic describes the release notes for AnalyticDB for MySQL and provides links to the relevant references.

Usage notes

Take note of the following items during minor version updates of AnalyticDB for MySQL clusters:

For AnalyticDB for MySQL clusters in reserved mode for Cluster Edition or AnalyticDB for MySQL clusters in elastic mode for Cluster Edition that have 32 cores or more, data read and write operations are not interrupted when engine versions are updated. Within 5 minutes before the update is complete, queries may encounter transient connections.
For AnalyticDB for MySQL clusters in elastic mode for Cluster Edition that have 8 or 16 cores, data write operations may be interrupted for 30 minutes when engine versions are updated. Within 5 minutes before the update is complete, queries may encounter transient connections.
Minor version updates of AnalyticDB for MySQL clusters do not affect database access, account management, database management, or IP address whitelist settings.
During a minor version update of an AnalyticDB for MySQL cluster, network jitters may occur and affect write and query operations. Make sure that your application is configured to automatically reconnect to the AnalyticDB for MySQL cluster.
During a minor version update of an AnalyticDB for MySQL cluster, the cluster may encounter transient connections. Make sure that your application is configured to automatically reconnect to the AnalyticDB for MySQL cluster.

If you do not need to update the minor version of your AnalyticDB for MySQL cluster or an error occurs during the update process, you can cancel the scheduled minor version update. You can cancel only the scheduled events of a minor version update. For more information, see the "Cancel scheduled events" section of the Manage O&M events topic.

Warning

If the minor version of your AnalyticDB for MySQL cluster is earlier than the latest minor version, Alibaba Cloud pushes a notification at an irregular interval to inform you that the cluster needs to be updated to the latest minor version. We recommend that you update the minor version of your AnalyticDB for MySQL cluster at the earliest opportunity within six months after you receive the notification. Otherwise, you shall assume all liabilities for risks such as service interruptions and data loss.

December 2024

V3.2.4

Category	Feature	Description	References
New feature	Machine learning prediction by using SQL	SQL can be used to quickly deploy behavior sequence transformer (BST) models to execute machine learning jobs, implementing data preprocessing, training, and prediction.	Use SQL to implement machine learning prediction

November 2024

Category	Feature	Description	References
New feature	Lake cache	The lake cache feature is supported to cache the Object Storage Service (OSS) objects that are frequently accessed on high-performance NVMe SSDs to improve the reading efficiency of OSS data.	Lake cache

October 2024

Category	Feature	Description	References
New feature	Backup and restoration	Data backup sets can be deleted and the data backup feature can be disabled in the AnalyticDB for MySQL console.	Manage backups
New feature	Zero-ETL	The zero-ETL feature is supported to synchronize Lindorm data. You can create data synchronization tasks from Lindorm to AnalyticDB for MySQL to synchronize and manage data in an end-to-end manner and integrate transaction processing with data analysis.	Import data from Lindorm

September 2024

Data Lakehouse Edition

Category	Feature	Description	References
New feature	Cross-region cluster cloning	Clusters can be cloned across regions.	Clone a cluster

V3.2.2

Category	Feature	Description	References
New feature	Batch creation of MaxCompute external tables	Multiple MaxCompute external tables can be created at a time.	IMPORT FOREIGN SCHEMA
	Support for aggregate functions in configuring incremental refresh for materialized views	The `MAX()`, `MIN()`, `APPROX_DISTINCT()`, `COUNT(DISTINCT)`, and `AVG()` functions can be included in the QUERY BODY parameter when you configure incremental refresh for materialized views.	Configure incremental refresh for materialized views (preview)
	Access to MaxCompute external tables in arrow API mode	The arrow API mode is supported to read and write MaxCompute external tables. Compared with the traditional tunnel mode, the arrow API mode can improve data access and processing efficiency.	Use external tables to import data to Data Lakehouse Edition
Optimized feature	FROM_UNIXTIME function	The FROM_UNIXTIME function can be used to convert a UNIX timestamp to the DATETIME format.	Date and time functions

August 2024

Category	Feature	Description	References
New feature	Selection of the Spark engine for interactive resource groups	The Spark engine can be selected when you create an interactive resource group in an AnalyticDB for MySQL Data Lakehouse Edition cluster. You can run only Spark jobs in interactive resource groups by using the Spark engine. Spark jobs are run in an interactive manner.	Create a resource group
New feature	Limits on the number of zero-ETL tasks	The number of zero-ETL tasks from ApsaraDB RDS for MySQL or PolarDB for MySQL to AnalyticDB for MySQL is limited.	Use zero-ETL to synchronize data from ApsaraDB RDS for MySQL Use zero-ETL to synchronize data from PolarDB for MySQL

July 2024

Category	Feature	Description	References
New feature	Release of the desired state of Basic Edition	The desired state of AnalyticDB for MySQL Basic Edition is released. The desired state of Basic Edition runs in single-replica mode and provides the same features as Enterprise Edition. Basic Edition uses a single-replica storage architecture and does not support high availability. Basic Edition is suitable for business scenarios that require low-cost hot data storage but do not require high availability.	Editions

V3.2.1

Category	Feature	Description	References
New feature	Next-generation storage engine	The next-generation storage engine `XUANWU_V2` is launched by AnalyticDB for MySQL. The engine caches cold data to the Enterprise SSDs (ESSDs) to speed up data reading and provides the next-generation column-oriented storage that supports higher I/O concurrency and reduces the memory usage. The engine also allows you to enable the compaction service to perform local compaction operations in an independent process by using an independent resource pool. This reduces resource usage and improves service stability.	XUANWU_V2 engine
	Incremental refresh for multi-table materialized views	Incremental refresh is supported for multi-table materialized views. The incremental data of multiple tables that are joined together can be automatically refreshed to the corresponding multi-table materialized view. This improves data query performance and data analysis efficiency.	Configure incremental refresh for materialized views (preview)
	Invocation of user-defined functions (UDFs) by using the REMOTE_CALL() function	The REMOTE_CALL() function can be used to invoke custom functions that you create in Function Compute (FC). This way, you can use UDFs in AnalyticDB for MySQL.	UDFs
	Forcible deletion of databases	The CASCADE keyword is supported in the DROP DATABASE statement to forcibly delete a database, including all tables in the database.	DROP DATABASE
	Wide table engine	The wide table engine is supported for Data Lakehouse Edition. The wide table engine is compatible with the capabilities and syntax of the open source columnar database ClickHouse and can handle large amounts of columnar data.	Wide table engine
	Path analysis functions	The SEQUENCE_MATCH() and SEQUENCE_COUNT() functions are supported to analyze user behavior and check whether the user behavior matches the specified pattern.	Path analysis functions
	SSL encryption	SSL encryption is supported to encrypt data transmitted between a Data Warehouse Edition cluster and a client. This prevents data from being listened to, intercepted, and tampered with by third parties.	SSL encryption
	Support for complex MaxCompute data types by MaxCompute external tables	Complex MaxCompute data types, such as ARRAY, MAP, and STRUCT, are supported for MaxCompute external tables of Data Lakehouse Edition clusters.	CREATE EXTERNAL TABLE
	Subscription to AnalyticDB for MySQL binary logs by using Flink	Realtime Compute for Apache Flink can be used to consume AnalyticDB for MySQL binary logs in real time.	Use Realtime Compute for Apache Flink to subscribe to AnalyticDB for MySQL binary logs
	Support for the ROARING BITMAP type by AnalyticDB for MySQL internal tables	The ROARING BITMAP type is supported.	Roaring bitmap functions
Optimized feature	Change of LIFECYCLE from a required keyword to an optional one	If you do not specify the LIFECYCLE keyword when you create a table, partition data is permanently retained.	CREATE TABLE
	Table-level partition lifecycle management	For AnalyticDB for MySQL clusters of V3.2.1.1 or later, the partition lifecycle is managed at the table level, but not the shard level. The LIFECYCLE n parameter specifies that up to n partitions can be retained in each table.	CREATE TABLE
	Import of OSS data to AnalyticDB for MySQL by using external tables	The absolute path name and the asterisk (*) wildcard are supported for the url parameter when you use external tables to import OSS data to AnalyticDB for MySQL.	Use external tables to import data to Data Warehouse Edition
	Automatic validity check of column names at table creation	Column names are automatically checked against naming conventions of AnalyticDB for MySQL when you execute the CREATE TABLE statement to create a table. If a column name does not meet the naming conventions, an error is returned. For information about the naming conventions of column names, see the "Naming limits" section of the Limits topic.	None

June 2024

Category

Feature

References

New feature

AnalyticDB for MySQL Enterprise Edition and Basic Edition are released.

Enterprise Edition runs in a cluster mode and is an integrated edition of Data Lakehouse Edition and Data Warehouse Edition that provides the same features as Data Lakehouse Edition. Enterprise Edition supports capabilities in elastic mode, such as resource group isolation, elastic resource scaling, and tiered storage of hot and cold data. Enterprise Edition also supports capabilities in reserved mode, such as high throughput, real-time writes and high-concurrency, real-time queries.
Basic Edition runs in standalone mode and supports tiered storage of hot and cold data. Basic Edition does not provide distributed capabilities, high availability, resource group isolation, or scheduled scaling. You cannot change a cluster from Basic Edition to Enterprise Edition.

Editions

April 2024

Category	Feature	Description	References
New feature	Query rewrite	The query rewrite feature of materialized views is supported. After you enable this feature, the optimizer determines whether a query can use pre-computed and stored data in materialized views. This way, the optimizer partially or entirely rewrites the original query to a query that can use materialized views.	Query rewrite of materialized views
	Synchronization of Simple Log Service (SLS) data by using data synchronization	The data synchronization feature can be used to synchronize data in real time from an SLS Logstore to an AnalyticDB for MySQL cluster based on a specific offset. This helps meet your business requirements for real-time analysis of log data.	Use data synchronization to synchronize data from Simple Log Service to Data Warehouse Edition Use data synchronization to synchronize data from Simple Log Service to Data Lakehouse Edition
	Zero-ETL	The zero-ETL feature is supported to help you synchronize and manage data, integrate transaction processing with data analysis, and focus on data analysis. You can create data synchronization tasks from ApsaraDB RDS for MySQL or PolarDB for MySQL to AnalyticDB for MySQL.	Use zero-ETL to synchronize data
	Time zone selection at cluster creation	The time zone parameter can be selected for an AnalyticDB for MySQL cluster at cluster creation based on your business requirements. After you select a time zone, the system performs time-related data writes based on the selected time zone.	Create a cluster
	Self-service minor version update	The minor version of a Data Warehouse Edition cluster can be viewed and updated in the AnalyticDB for MySQL console.	Update the minor version of a cluster
	Vertical scaling of reserved storage resource specifications	Reserved storage resource specifications can be scaled up or down for Data Lakehouse Edition clusters.	Scale a Data Lakehouse Edition cluster
	Use of a Spark distributed SQL engine in DataWorks	A Spark distributed SQL engine of AnalyticDB for MySQL Data Lakehouse Edition can be registered as an execution engine by registering Cloudera's Distribution Including Apache Hadoop (CDH) clusters to DataWorks. This way, you can develop and run Spark SQL jobs in DataWorks.	Use a Spark distributed SQL engine in DataWorks
	Display of the progress bar in creation or configuration change of a cluster	A progress bar is displayed when you create or change the configurations of a Data Warehouse Edition cluster.	Create a cluster

March 2024

Data Lakehouse Edition

Category	Feature	Description	References
New feature	Spot instance	The spot instance feature can be enabled for job resource groups in Data Lakehouse Edition clusters. After you enable the spot instance feature for a job resource group, Spark jobs that run in the resource group attempt to use the spot instance resources. Compared with AnalyticDB compute unit (ACU) elastic resources, spot instance resources help you significantly reduce the costs of Spark jobs.	Spot instances

February 2024

Category	Feature	Description	References
New feature	Intelligent assistant	An intelligent assistant is provided in the AnalyticDB for MySQL console. The intelligent assistant answers your questions and helps you quickly resolve issues. Note The intelligent assistant supports only the Chinese language.	None
	Spark distributed SQL engines	AnalyticDB for MySQL Data Lakehouse Edition Spark provides managed services for open source Spark distributed SQL engines to develop Spark SQL jobs. This helps you easily analyze, process, and query data to improve SQL efficiency.	Use a Spark distributed SQL engine to develop Spark SQL jobs
	Access to OSS-HDFS	AnalyticDB for MySQL Data Lakehouse Edition Spark can be used to access OSS-HDFS.	Access OSS-HDFS
	Storage overview	The data size of a cluster or a table can be viewed on the Storage Overview page of the AnalyticDB for MySQL console.	Storage analysis

V3.1.10

Category	Feature	Description	References
New feature	Primary and foreign key constraints	Primary and foreign key constraints can be used to eliminate unnecessary joins to improve database query performance.	Use primary and foreign key constraints to eliminate unnecessary joins
	Monthly execution of resource scaling plans	Resource scaling plans can be configured to execute every month in Data Warehouse Edition.	Create a resource scaling plan
	Multi-cluster scaling models	The multi-cluster feature can be enabled for resource groups in Data Lakehouse Edition. A multi-cluster scaling model allows AnalyticDB for MySQL to automatically scale resources based on query loads to meet resource isolation and high concurrency requirements for resource groups.	Multi-cluster scaling models
	Variable-length binary functions	The AES_DECRYPT_MY() and AES_ENCRYPT_MY() functions are supported.	Variable-length binary functions
	JSON functions	The JSON_REMOVE() function is supported.	JSON functions
	Plan cache	The plan cache feature is supported to allow you to cache execution plans of SQL statements. When you execute SQL statements that share the same SQL pattern, AnalyticDB for MySQL uses the cached execution plan of the SQL pattern to accelerate SQL compilation optimization and improve query performance.	Plan cache
	Elastic import	The elastic data import method is supported for Data Lakehouse Edition. Elastic import consumes a small amount of storage resources or does not consume computing and storage resources. This reduces impacts on real-time data reads and writes and improves resource isolation.	Data import methods
	Asynchronous scheduling of extract, transform, load (ETL) tasks by using Data Management (DMS)	The task orchestration feature of DMS can be used to asynchronously schedule ETL tasks.	None
	Modification of workload management rules	The WLM syntax can be used to modify workload management rules.	WLM
Optimized feature	Basic statistics	The collection policy for basic statistics is optimized.	None
	Column group statistics	The collection policy for column group statistics is optimized.	None
	Internal Error error message	The Internal Error error messages is optimized to help you quickly identify issues.	None
	Asynchronous generation of splits	For external tables that have large amounts of data, AnalyticDB for MySQL can asynchronously generate splits to reduce the amount of time required to generate execution plans.	None
	Split flow control	The split flow control feature for scanning OSS and MaxCompute external tables is optimized.	None
	Parameter check policy for making RC HTTP calls	The parameter check policy for making RC HTTP calls is optimized to prevent SQL injections.	None
	Memory usage of storage nodes	The memory usage of storage nodes is optimized to reduce garbage collection (GC) frequency and improve system stability.	None
Fixed issue	Materialized views	The following issue is fixed: An error is returned for the ARRAY_AGG() function when you use the CREATE VIEW statement to create a view.	None
	On-premises data import by using the LOAD DATA statement	The following issue is fixed: When you use the LOAD DATA statement to import on-premises data to Data Warehouse Edition, CSV files are incompatible or data is disordered.	None
	Storage of cold data	The cold data storage issue is fixed to improve the query hit ratio and query performance.	None

November 2023

Data Warehouse Edition

Category	Feature	Description	References
New feature	Diagnostics	The diagnostics feature is supported. This feature allows you to diagnose the running status of clusters within a specific period of time. AnalyticDB for MySQL performs joint analysis based on monitoring data, log data, and the status of databases and tables. AnalyticDB for MySQL evaluates the health status of clusters from multiple aspects, such as resource usage, workload changes, SQL queries, operators, and storage, to help you efficiently identify and resolve issues.	Diagnostics
New feature	Change of virtual private clouds (VPCs) and vSwitches	VPCs and vSwitches can be changed.	Change the VPC and vSwitch of a cluster

Data Lakehouse Edition

Category	Feature	Description	References
New feature	Custom Spark images	Custom Spark images are supported. If the default image of AnalyticDB for MySQL Spark cannot meet your business requirements, you can add the software packages and dependencies required for Spark jobs to the default image to create a custom image. When you develop Spark jobs, you can specify the custom image as the execution environment.	Custom Spark images
New feature	Development of interactive Jupyter jobs	A Docker image can be used to start the interactive JupyterLab development environment. This environment helps you connect to AnalyticDB for MySQL Spark and perform interactive testing and computing based on elastic resources.	Develop an interactive Jupyter job

October 2023

V3.1.9

Category	Feature	Description	References
New feature	Common table expression (CTE) execution optimization	If a CTE subquery is referenced repeatedly, the subquery can be executed only once to improve query performance. By default, this feature is disabled. You can enable this feature by using the cte_execution_mode parameter.	WITH
	Access to Hudi data by using XIHE SQL	XIHE SQL can be used to access Hudi data from Data Lakehouse Edition.	OSS external tables
	MV_PROPERTIES configuration	An elastic resource group can be specified to create and refresh materialized views to improve query efficiency.	Elastic materialized views
	Column group statistics	The statistics on multiple columns of a table can be collected to describe how these columns correlate with each other.	Statistics
	Manual connection of partition statistics	The `ANALYZE TABLE` statement can be used for Data Lakehouse Edition to collect statistics on partitions of OSS external tables.	Collect statistics on partitions
	Variable-length binary functions	The `ZIP()`, `UNZIP()`, `GZIP()`, and `GUNZIP()` functions are supported.	Variable-length binary functions
	Incremental refresh for materialized views	Incremental refresh can be configured when you create a materialized view.	Configure incremental refresh for materialized views (preview)
	Forcible overwriting of existing properties of workload management rules	After you use the WLM syntax to create a workload management rule, existing properties of the rule can be forcibly overwritten.	WLM
	AI_GENERATE_TEXT() function	The AI_GENERATE_TEXT() function can be used to analyze unstructured data and generate structured data in Data Warehouse Edition.	None
	Multi-statement	Multiple SQL statements that are separated by semicolons (;) can be consecutively executed. By default, the multi-statement feature is disabled. You can execute the `SET ADB_CONFIG ALLOW_MULTI_QUERIES=true;` statement to enable the feature.	None
	Statistics on Hive external tables	The number of rows in ORC external tables can be collected in real time to optimize complex queries of ORC external tables.	None
Optimized feature	JOIN optimization	The filter scenarios and data transfer efficiency are optimized when a hash join is used to join tables. A small table can be used in a subquery and efficiently filtered to transfer data that meets requirements to the main query.	None
	Optimization of vectorized reading of Parquet files	The query efficiency of Parquet files is improved.	None
	Optimization of Aggregation operators	The execution efficiency of Aggregation operators is optimized in GROUP BY scenarios that use a STRING-type column or multiple columns.	None
	Optimization of dictionary encoding	Dictionary encoding is used to improve the performance of GROUP BY operations.	None
	Analyzer optimization	The method of specifying custom dictionaries for the IK analyzer is optimized.	None
	Optimization of executor nodes	The startup speed of executor nodes in job resource groups is increased.	None
	Optimization of INSERT OVERWRITE	An external table can be used in multiple INSERT OVERWRITE operations at the same time.	None
	Optimization of asynchronous jobs	The maximum length of the result set of asynchronous queries is increased.	None
Fixed issue	Precision of the DECIMAL type	The following issue is fixed: Row-oriented engines do not support the precision change of the DECIMAL type.	None
	Statistics on Hive external tables	The following issue is fixed: An extended amount of time is required to collect information about Hive external tables.	None
	WITH	The following issue is fixed: The table alias that is enclosed in grave accents (``) cannot be identified in the WITH clause.	None
	File names of external tables	The following issue is fixed: An error message is returned if the file name of an external table contains a colon (:).	None

September 2023

Data Lakehouse Edition

Category	Feature	Description	References
New feature	Spark application performance diagnostics	The Spark application performance diagnostics feature helps you quickly locate and analyze performance bottlenecks to resolve issues.	Spark application performance diagnostics
	Public network configuration for Spark application access	A public network can be configured for Spark applications to access self-managed databases or third-party cloud services.	Public network configuration for Spark application access
	Access to MySQL data by using Spark SQL	Spark SQL can be used to access self-managed MySQL databases or Alibaba Cloud MySQL databases.	Read MySQL data
	Access to Lindorm data by using Spark SQL	Spark SQL can be used to access Hive tables and wide tables of Lindorm.	Read Lindorm data

June 2023

Category

Feature

Description

References

New feature

Resource overview and job usage statistics

The following information about Data Lakehouse Edition cluster resources is displayed in the AnalyticDB for MySQL console:

Cluster information: the amounts of computing and storage resources, including the reserved and elastic resources.
Resource group information: the total amount of computing resources, the amount of reserved computing resources, and the maximum amount of computing resources.
Job information: the total amount of computing resources, the amount of reserved computing resources, and the amount of elastic computing resources.

View monitoring information about resource groups

Optimized feature

Change of the default data backup cycle

The default data backup cycle of Data Warehouse Edition clusters is changed from at least twice a week to at least once a week.

Manage backups

May 2023

V3.1.7 to V3.1.8

Category	Feature	Description	References
New feature	Improved monitoring and alerting	The instance health status and cluster health status metrics are supported.	View monitoring information of AnalyticDB for MySQL
	Priority queues and concurrency control of interactive resource groups	The priority queue feature is supported for queries in interactive resource groups. You can configure query priorities to allow queries to enter one of the following priority queues: LOWEST, LOW, NORMAL, and HIGH. You can also configure the number of concurrent queries for queues.	Priority queue and concurrency of interactive resource groups
	Priority queues of job resource groups	The priority queue feature is supported for jobs in job resource groups. You can configure job priorities to allow jobs to enter one of the following priority queues: LOWEST, LOW, NORMAL, and HIGH. Jobs that have higher priorities are preferentially run.	Priority queues of job resource groups
	Precision change of the DECIMAL type	The precision of the DECIMAL type can be changed from low to high.	ALTER TABLE
	Change of the data type	You can change an integer type such as TINYINT, SMALLINT, INT, BIGINT, SHORT, and LONG to a floating-point type such as FLOAT and DOUBLE, or the DECIMAL type.
	ALTER TABLE PARTITION	The partition function of a table can be changed.
Optimized feature	Optimizer optimization: Eager aggregation and automatic two-phase aggregation rules are supported. By default, the Cascades optimizer is enabled. The Swap Outer Join rule is supported. The executor restart or upgrade upon a task failure does not affect the running of tasks. The write performance of the INSERT OVERWRITE statement is improved. The restart of storage nodes is accelerated. JSON optimization: The issue about the `IS NOT NULL or IS NULL` operator of the JSON_EXTRACT() function is fixed. The performance compromise issue of the C-Store storage engine due to pushdown failure of the JSON_ARRAY() function is fixed. The performance of data scan operators is improved.		None

April 2023

Data Lakehouse Edition

Category	Feature		References
New feature	ACU-hour plans can be purchased to offset the amount of reserved computing resources, reserved storage resources, and elastic resources of pay-as-you-go clusters, and elastic resources of subscription clusters.		ACU-hour plans

February 2023

V3.1.6.4

Category	Feature	Description	References
New feature	Roaring bitmap functions	Roaring bitmaps are efficiently compressed bitmaps that are widely used in various programming languages and big data platforms for deduplication, tag-based filtering, and computing of time series data.	Roaring bitmap functions
	Funnel analysis functions	Funnel analysis is a common type of conversion analysis. It is used to reflect the conversion rates of user behavior in various stages of a process. The following functions are supported: WINDOW_FUNNEL(), RETENTION(), RANGE_RETENTION_COUNT(), and RANGE_RETENTION_SUM().	Funnel and retention functions
	UPDATE JOIN	The UPDATE statement can be used together with JOIN to update the data of multiple tables.	Update multiple tables
	ApsaraDB RDS for MySQL, ApsaraDB for MongoDB, MaxCompute, OSS, and Tablestore external tables	External tables can be used to import data from ApsaraDB RDS for MySQL or ApsaraDB for MongoDB to AnalyticDB for MySQL Data Lakehouse Edition. External tables can be used to import data from MaxCompute to AnalyticDB for MySQL Data Lakehouse Edition. For Data Lakehouse Edition clusters, partitioned OSS external tables provide partition mapping and support a variety of formats such as CSV, JSON, Parquet, ORC, and Avro. The formats can be used to import data that is stored in OSS and contains more than 100,000 partitions. The Tablestore connector is supported to import data from Tablestore.	Use external tables to import data to Data Lakehouse Edition Import data from ApsaraDB for MongoDB Use external tables to import data to Data Lakehouse Edition Use external tables to import data to Data Lakehouse Edition Import data from Tablestore
	Character sets supported for MySQL external tables	MySQL character sets can be specified by using the `charset` parameter when you create an ApsaraDB RDS for MySQL or self-managed MySQL external table.	CREATE EXTERNAL TABLE
	Cost-based optimizer (CBO) update	The automatic statistics collection feature is supported in Data Warehouse Edition. Statistics about data columns can be used to help the query optimizer generate high-quality execution plans.	Statistics
	Intelligent workload management	The workload management feature separates and throttles queries by assigning them to queues of different priorities. You can customize rules to intercept bad queries and assign queries to queues.	Overview Configure workload management rules WLM Best practices for workload management
Optimized feature	New import models are supported to improve import performance for newly created tables. The hash and sort performance of Window operators is optimized to support adaptive aggregation. Jetty is replaced with Netty to reduce network connections and CPU consumption. By default, Netty is enabled for control links. You can enable Netty for data links. The types and value ranges of partition fields can be specified in DDL statements. This optimizes the partition pruning performance of OSS external tables. The memory model used for importing data based on OSS and MaxCompute external tables is optimized. The performance of reading external tables is also optimized. CacheFS is optimized to reduce caching for hot and cold data and improve stability.		None

V3.1.5.8

Category	Feature	Description	References
New feature	Full-text search	The following analyzers are built into AnalyticDB for MySQL to implement full-text search: Standard, Ngram, Edge_ngram, and Pattern.	Analyzers for full-text indexes

V3.1.5.10

Category	Feature	Description	References
New feature	Regular expression functions	The following regular expression functions are supported: `REGEXP_MATCHES()`, `REGEXP_SUBSTR()`, `REGEXP_INSTR()`, and `REGEXP_REPLACE()`.	Regular functions

January 2023

Data Warehouse Edition

Category	Feature	References
New feature	The SQL diagnostics feature is supported. This feature allows you to view stage and task details to improve analysis efficiency of slow queries.	Use stage and task details to analyze queries
	The performance level of ESSDs can be changed.	Pricing for Data Warehouse Edition Scale a Data Warehouse Edition cluster

Data Lakehouse Edition

Category	Feature		References
New feature	AnalyticDB for MySQL Data Lakehouse Edition is available. Besides the real-time analysis capability of Data Warehouse Edition, Data Lakehouse Edition provides the batch processing capability.		Editions Getting started with Data Lakehouse Edition

November 2022

Data Warehouse Edition

Category	Feature		References
New feature	AnalyticDB for MySQL Data Warehouse Edition is available in the Philippines (Manila) and Thailand (Bangkok) regions.		Pricing for Data Warehouse Edition

August 2022

V3.1.5.0

Category	Feature	Description	References
New feature	Enhancement of the DECIMAL type	Decimal numbers can be converted to a higher precision, and variable-length decimal numbers are supported. This feature improves the I/O efficiency of DECIMAL data.	None
	Table-level throttling	The writing rate of DML statements is limited for specific tables to ensure overall performance. By default, table-level throttling is disabled.	None
	Memory management on wide tables	The memory management on wide tables is optimized to reduce the consumption of memory resources.	None
	JSON_UNQUOTE() function	The JSON_UNQUOTE() function can be used to unquote the value specified by `json_value`, escape specific characters in `json_value`, and then return the processing result.	JSON_UNQUOTE
	JSON_CONTAINS() function	The JSON_CONTAINS() function can be used to determine whether a given candidate is contained within a JSON document or whether the candidate exists in a specified path within the JSON document.	JSON_CONTAINS
	JSON_CONTAINS_PATH() function	The JSON_CONTAINS_PATH() function can be used to determine whether a specified path exists in the given JSON document.	JSON_CONTAINS_PATH
Optimized feature	TopN operator: In scenarios where the `ROW_NUMBER()` function is used to sort data and the WHERE clause is used to implement semantics of TopN, a variety of measures such as sorting, condition pushdown, and pre-aggregation can be taken to reduce the amount of data that is involved in computing and transmission. The performance of TopN operators can be improved tenfold while memory consumption is reduced by 90%. Window operator: General-use window functions are optimized to improve performance by three to seven times. Different algorithms are used based on different data features such as data aggregation degree and sequence ordering. Partial Agg operator: Self-adaption capabilities of two-phase aggregation are improved. Data features of data aggregation are dynamically collected during execution to determine whether pre-aggregation is required. The performance of aggregation operators can be improved by two to four times if no aggregation degree exists. System logs: When the disk usage exceeds the threshold, logs can be automatically cleared.		None

March 2022

Category	Feature	Description	References
New feature	Schema optimization	The schema optimization feature is supported to provide optimization suggestions of the hot and cold data optimization, index optimization, and distribution key optimization types based on intelligent statistic analysis. This feature can help reduce costs and improve efficiency for the use of AnalyticDB for MySQL clusters.	Schema optimization

December 2021

V3.1.4.13 to V3.1.4.16

Category	Feature	References
New feature	Two data replicas and one log replica are configured based on the Raft algorithm to ensure data reliability and reduce storage overheads.	None
	High availability is supported for the nameservice of Apsara File Storage for HDFS when data is exported to HDFS.	Export data to Apsara File Storage for HDFS
Optimized feature	The lock granularity for CREATE VIEW is optimized to improve concurrency performance. The upload speed of full backup files is increased during specification changes. The index creation performance for complex data types is improved. The memory management of background tasks is optimized.	None

September 2021

V3.1.4.12

Category	Feature		References
Optimized feature	The performance of the hash join algorithm to create a hash table is improved.		None

August 2021

V3.1.4.11

Category	Feature	Description	References
New feature	API operations related to cluster running reports	API operations can be called to query metrics in a cluster running report.	DescribeTableAccessCount DescribeSqlPattern
Optimized feature	The download speed of full data is increased. Inappropriate index condition pushdown (ICP) is blocked. By default, 128 indexes are not pushed down. This configuration can be changed.		None

V3.1.4.10

Category	Feature	Description	References
New feature	O&M event management	The database upgrade time can be viewed and adjusted in the AnalyticDB for MySQL console.	Manage O&M events
Optimized feature	The TIME() function can be pushed down. The table scan performance is improved. Mutual exclusion and fair scheduling are supported for data import.		None

July 2021

V3.1.4.9

Category	Feature	Description	References
New feature	Data import to and export from Apsara File Storage for HDFS by using external tables	External tables can be used to import Apsara File Storage for HDFS data to AnalyticDB for MySQL and export AnalyticDB for MySQL data to Apsara File Storage for HDFS.	Use external tables to import data to Data Warehouse Edition Export data to Apsara File Storage for HDFS
	SQL diagnostics	The details of SQL queries can be viewed and filtered based on categories such as the top 100 most time-consuming queries and queries that failed to be executed. Also, SQL queries can be optimized based on diagnostic results and optimization suggestions.	Overview
	End-to-end data management	An end-to-end data management portal is added to the AnalyticDB for MySQL console. Data assets can be managed and jobs can be developed and scheduled by using DMS.	Manage data assets Schedule XIHE SQL jobs by using DMS
	Custom analyzers and dictionaries in full-text search scenarios	Custom analyzers and dictionaries can be configured in full-text search scenarios.	Analyzers for full-text indexes
Optimized feature	Connections to frontend nodes are optimized from single-threaded connections to multi-threaded connections. This improves write performance in a linear manner. Performance is improved when the `TRUNCATE TABLE` statement is frequently executed. By default, the `REPLACE INTO` statement is atomic to prevent sudden changes to data.		None

March 2021

V3.1.1.9 to V3.1.3.9

Category	Feature	Description	References
New feature	Computing resource grouping	Computing resources of AnalyticDB for MySQL clusters in elastic mode can be divided into resource groups for isolation.	Create a resource group
	Tiered storage of hot and cold data	Table data of AnalyticDB for MySQL clusters in elastic mode can be defined as hot or cold data. You can switch between hot and cold storage.	Tiered storage of hot and cold data
	Cluster mode change	AnalyticDB for MySQL clusters can be changed from reserved mode to elastic mode.	None
	Compatibility with time formats in AnalyticDB for MySQL	Time formats in AnalyticDB for MySQL V2.0 are supported. Example: 2020-08-03T23:59:59.	None
	Index creation or deletion for JSON fields by executing the ALTER TABLE statement	Indexes for JSON fields can be disabled by executing the ALTER TABLE statement.	ALTER TABLE
	BINARY type	The BINARY type is supported for the metadata of the protocol layer.	None
	Export of file headers during export from AnalyticDB for MySQL to a single OSS object	File headers can be exported when you export data from AnalyticDB for MySQL to a single OSS object by using an external table.	Export data to OSS
	Maximum number of rows that can be generated in an object when you export data from AnalyticDB for MySQL to OSS by using an external table	If the number of exported rows exceeds the maximum number, extra rows are exported to one or more new objects. You can specify both the maximum size and maximum number of rows in an object. Written data that first triggers the limit is exported to a new object.	None
	SQL plan module	Execution plans of slow SQL queries can be viewed in the AnalyticDB for MySQL console.	Query process and execution plan
	`INSERT INTO SELECT ON DUPLICATE KEY UPDATE`	This query is supported when the input values in the UPDATE column are constants or when the input values in the UPDATE column are those in the SELECT column.	None
	File format of OSS external tables	The ORC format is supported for OSS external tables.	None
	Priority of the BATCH LOAD statement	Hints can be used to specify the priority of the BATCH LOAD statement.	None
Optimized feature	Performance of the LIMIT n clause	Performance is improved when you use the pushdown logic of the LIMIT n clause to filter data.	None
	Compatibility	The table creation statement is compatible with the BOOLEAN type.	None
	Database naming conventions	Database names can start with an uppercase letter or underscore (_).	None

July 2020

V3.1.1.6

Category	Feature	Description	References
New feature	Timestamp and Datetime columns	When the MODIFY COLUMN statement is executed to modify a Timestamp or Datetime column, the ON UPDATE CURRENT_TIMESTAMP clause is supported.	None
	Table and column naming conventions	Table and column names support Chinese characters.	None
	Requirements for creating an OSS external table	The following requirements must be met when you create an OSS external table: The partition key columns must be specified at the end of the field list. Otherwise, the table cannot be created. The first row in an OSS object can be set as a file header. The system skips the first row when the system reads data from the object.	Use external tables to import data to Data Warehouse Edition
	`CREATE TABLE AS SELECT`	The `CREATE TABLE AS SELECT` statement can be executed to create a table.	CREATE TABLE AS SELECT (CTAS)
Optimized feature	Fields of the BOOLEAN type	The default values for the fields of the BOOLEAN type can be 0 or 1.	None
Optimized feature	SHOW DATABASES	The permissions to list databases can be granted when the SHOW DATABASES statement is executed.	None

April 2020

V3.0.9.6

The following database software upgrades are performed for users of AnalyticDB for MySQL Basic Edition to improve service quality.

Category	Feature	Description	References
New feature	Geometry functions	Geometry functions are supported.	Operation functions
	JSON_EXTRACT() function	The JSON_EXTRACT() function is supported.	JSON_EXTRACT
	INSERT INTO VALUES(FROM_UNIXTIME(...))	The `INSERT INTO VALUES(FROM_UNIXTIME(...))` statement is supported.	None
	Nested-loop join (NLJ)	NLJ is supported for data join.	None
	Power BI connection	Power BI can be connected to the protocol layer.	None
	Database naming conventions	Hyphens (-) can be included in database names. Note Hyphens (-) must be enclosed in grave accents (``).	None
Optimized feature	Zero dates	Zero dates (0000-00-00) are converted to NULL.	None
	DIV() function of the DECIMAL type	The DIV() function of the DECIMAL type is supported as in MySQL.	DIV
	CAST() function of the JSON type	The CAST() function is supported for JSON data as in MySQL and Apache Hive.	JSON
	Slow query logging threshold	The slow query logging threshold is set to 1 second.	None

March 2020

V3.0.9

Category	Feature	Description	References
New feature	JSON data types and related JSON functions	Complex JSON data types and related JSON functions are supported.	JSON JSON functions
New feature	`SELECT NOW()`	`The SELECT NOW() statement is supported`.	None
Optimized feature	Maximum number of tables	The maximum number of tables that can be created in a Cluster Edition cluster of the minimum specifications is increased from 512 to 800. The minimum specifications indicate that the Cluster Edition cluster has only two node groups.	None
	Compatibility with DDL statements	Compatibility with DDL statements in AnalyticDB for MySQL V2.0 is improved to enable smooth data migration to AnalyticDB for MySQL V3.0 clusters. Your business is not affected during data migration.	None
	Compatibility with business intelligence (BI) tools	AnalyticDB for MySQL V3.0 improves compatibility with BI tools and is fully compatible with Power BI.	Power BI

February 2020

V3.0.8

Category	Feature	Description	References
New feature	MariaDB JDBC Connector	MariaDB Java Database Connectivity (JDBC) Connector is supported.	None
	Specifications applicable to Cluster Edition	The storage-intensive specification S8 is added for AnalyticDB for MySQL Cluster Edition clusters. S8 is ideal for scenarios that do not require high concurrency and performance.	None
	Flexible purchase of clusters	Node groups can be purchased and scaled out in pairs. This allows you to purchase clusters on demand and reduces costs.	None
	Availability in Alibaba Finance Cloud	AnalyticDB for MySQL is available in the China East 1 Finance, China East 2 Finance, and China South 1 Finance regions of Alibaba Finance Cloud.	None
	Availability on the international site (alibabacloud.com)	AnalyticDB for MySQL is available in the China (Hong Kong), Indonesia (Jakarta), and Malaysia (Kuala Lumpur) regions.	None
Optimized feature	Time types	The TIMESTAMP and DATETIME data types are compatible with the NO_ZERO_DATE mode of MySQL SQL_MODE.	None

December 2019

V3.0.7

Category	Feature	Description	References
New feature	Specification C24	The compute-intensive specification C24 is added for AnalyticDB for MySQL clusters. C24 is ideal for scenarios that require sophisticated computing capabilities.	None
	Configuration upgrade	Cluster specifications can be upgraded. You can perform switchovers between two of the following specifications within seconds: C8, C4, and C24.	None
	Monitoring and alerting	The monitoring and alerting feature is supported. You can use CloudMonitor to set thresholds for all metrics. An alert is triggered when a threshold is reached.	Configure an alert rule
	Query termination	The query termination feature is supported. You can view and terminate running queries in real time in the AnalyticDB for MySQL console.	None
	Data synchronization from PolarDB-X to AnalyticDB for MySQL	Data Transmission Service (DTS) can be used to synchronize data from PolarDB-X to AnalyticDB for MySQL in real time for data analytics.	None
	Availability on the international site (alibabacloud.com)	AnalyticDB for MySQL is released for international use. This service is available in the Singapore and Japan (Tokyo) regions.	None
Optimized feature	View creation	Window functions can be used to create views.	None
Optimized feature	Use scenarios of CTEs	CTEs can be used in the INSERT SELECT FROM clause.	INSERT SELECT FROM

September 2019

V3.0.6

Category	Feature	Description	References
New feature	Specification C4	The specification C4 is added to simplify the use of AnalyticDB for MySQL. We recommend that you use this specification in learning.	None
New feature	COLLECT_SET() function	The COLLECT_SET() function is supported.	None
Optimized feature	Creation and scaling time for clusters	The amount of time spent on creating and scaling clusters is shortened to reduce costs.	None

August 2019

V3.0.5

Category	Feature	Description	References
New feature	Default column value	The default value of a column can be set to the current time. Example: `gmt_create datetime NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP`.	None
	Oracle GoldenGate (OGG)	OGG is supported in AnalyticDB for MySQL to enhance data synchronization from Oracle to AnalyticDB for MySQL.	None
	Disk resizing	Flexible disk resizing is supported. This allows you to resize disks on demand and reduces costs.	None
	Availability in Alibaba Finance Cloud	AnalyticDB for MySQL is available in Alibaba Finance Cloud.	None
	Virtual e-commerce logistics platforms and CloudTmall	AnalyticDB for MySQL is available in virtual e-commerce logistics platforms and CloudTmall.	None
Optimized feature	Error message returned for modifying non-auto-increment keys	The error message that is returned when you change non-auto-increment keys to auto-increment keys is optimized to facilitate your understanding.	None

July 2019

V3.0.4

Category	Feature	Description	References
New feature	Backup and restoration	The backup and restoration feature is supported. You can restore data from backup sets to a point in time to maximize data restorability.	None
	LOAD DATA	The LOAD DATA LOCAL INFILE statement is supported.	LOAD DATA LOCAL INFILE
	Flexible purchase of services	Node groups can be purchased in pairs. For example, you can set Node Groups to 2, 4, 6, or 8 on the AnalyticDB for MySQL buy page.	None
	Data types and important functions	New data types and specific important functions are supported.	None
Optimized feature	Compatibility	AnalyticDB for MySQL is fully compatible with Navicat, FineReport, and FineBI, and its compatibility with Sequel Pro is greatly improved.	None