For installation and running instructions, see [Get started](https://docs.risingwave.com/docs/current/get-started/).
Main changes
SQL features

Query syntax:

Technical preview: Supports ASOF JOIN to join and find the closest matching record by the event time or another ordered property. feat(streaming): plan asof join risingwave#18683
Supports AGGREGATE: prefixed scalar function in streaming aggregation. feat(agg): support aggregate: prefixed scalar function in streaming agg risingwave#18205
Supports using user-defined aggregate functions as window function. fix(udf): allow udaf as window function risingwave#18181
Supports blocking subscription cursors and configuring cursor timeouts. feat(subscription): support blocking cursor risingwave#18675


SQL commands:

Enhances observability of cursors and subscription cursors by improving output results of SHOW SUBSCRIPTION CURSORS  and SHOW CURSORS commands. feat: improve observability for subscription/cursor SHOW commands risingwave#18896


SQL functions & operators:

Technical Preview: Supports the TVF postgres_query . feat(batch): support basic postgres tvf risingwave#18811


System catalog:

Supports pg_catalog.pg_sequence. feat: support pg_catalog.pg_sequence risingwave#18714
Supports information_schema.table_constraints. feat: support information_schema.table_constraints risingwave#18716
Supports information_schema.schemata. feat: support information_schema.schemata risingwave#18709



Connectors

Breaking change: Changes scan.startup.mode=latest for NATS source connector to start consuming from next available message instead of last one. fix(nats): align the semantics of scan.startup.mode=latest of NATS with others risingwave#18733
Public preview: Supports shared Kafka sources, which can be disabled by session variable  streaming_use_shared_source. feat: enable shared source in session variable by default, and add cluster-level config to disable risingwave#18749
Supports recursively scanning file sources. feat(source): configure whether to recursively scan file source risingwave#18324
Supports schemaless ingestion for data in JSON format from Kafka sources by using INCLUDE payload clause. feat: schemaless ingestion for encode json (INCLUDE payload) risingwave#18437
Adds a set of options for NATS source connector based on the async_nats crate. feat(source): add NATS source consumer parameters risingwave#17615
Adds a required option, consumer.durable_name, for NATS source connector. refactor: impl ack and migrate to durable consumer for Nats risingwave#18873
Supports option max_packet_size for MQTT sources. feat(mqtt): Add max_packet_size option to MQTT Source risingwave#18520
Supports option database.encrypt for SQL Server CDC sources. feat(mssql-cdc): add ssl support for sqlserver-cdc risingwave#18912
Supports ingesting data from a partitioned table for PostgreSQL CDC sources. fix(pg-cdc): fix support for partitioned table risingwave#18456
Supports option auto.schema.change for PostgreSQL CDC sources to enable replicating Postgres table schema change. feat(postgres-cdc): support replicating Postgres schema change risingwave#18760
Requires upstream table name to also be prefixed with database name when creating a SQL Server CDC table. fix(cdc): fix the filter expression of CdcFilter executor risingwave#18868
Adds JSON encode for file sinks, allowing users to sink JSON files into object storage. feat(sink): add json encode for file sink risingwave#18744
Supports create_table_if_not_exists option for Iceberg sink connector. feat(iceberg): support iceberg sink create table risingwave#18362
Supports WebHDFS sinks. feat(connector): add webhdfs sink risingwave#18293
Removes option bulk_write_max_entries for MongoDB sink and option default_max_batch_rows for DynamoDB sink. Adds options max_batch_item_nums and max_future_send_nums for DynamoDB sink. feat(sink): support async for mongodb  dynamodb risingwave#17645
Sets sink decoupling as the default policy for MongoDB, DynamoDB, and Redis sink connectors. feat(sink): support async for mongodb  dynamodb risingwave#17645
Supports option routing_column for ElasticSearch sinks, allowing a column to be set as a routing key. feat(sink): add es dynamic route risingwave#18698
Supports specifying batching strategy when sinking data in Parquet format. feat(sink): introduce batching strategy for  file sink risingwave#18472

Installation and deployment

Adds a CLI argument of --license-key-path for the meta node, enabling a background task to watch and reload license key from the specified file. feat(meta): watch and reload license key from file risingwave#18768

Cluster configuration changes

When visibility_mode is set to all, the latest uncommitted data will be queried, but consistency is no longer guaranteed between the tables. feat: deprecate current epoch in pinned snapshot risingwave#18230
Supports SET TIME ZONE INTERVAL '+00:00' HOUR TO MINUTE as equivalent to SET TIME ZONE UTC. feat: support SET TIME ZONE INTERVAL '+00:00' HOUR TO MINUTE; risingwave#18705
The etcd metastore is fully deprecated and unsupported. Users previously utilizing etcd metastore must manually migrate to a SQL backend (PostgreSQL, MySQL, or SQLite) to upgrade to v2.1.0

Full Changelog: risingwavelabs/risingwave@v2.0.4...v2.1.0