v1.9.1-rc.1

For installation and running instructions, see Get started.

Main changes

SQL features

• Query syntax:
  • Supports non-append-only temporal joins, where the outer side is not required to be append-only. #16286
• SQL commands:
  • Supports DISCARD ALL command for Npgsql compatibility. #16432
  • Supports creating, dropping, altering, and showing subscriptions. #14831
  • Supports cursors for subscription queues. #15180
  • Supports altering stream rate limit for sources and tables with a source. #16399
  • Supports RECOVER command to trigger an ad-hoc recovery. #16259
• SQL functions & operators:
  • Supports using a single ASCII character as an escape character with the LIKE clause. #16057
  • Supports jsonb_populate_record() and jsonb_populate_recordset(). #13421
• System catalog:
  • Supports system table rw_iceberg_files for displaying the files of an Iceberg source or table. #16180
  • Supports system table rw_iceberg_snapshot for listing all snapshots. #16175

Connectors

• Provides stable support for SQLAlchemy 2.0. risingwavelabs/sqlalchemy-risingwave#29
• Deprecates s3 connector. #16337
• Supports generated columns for non-shared CDC tables. #16522
• Supports time travel for Iceberg sources. #15866
• Blocks sink creation until backfill is completed by default. #16249
• Supports Kafka connector parameter properties.request.required.acks. #16482
• Adds connector parameter ssl.mode for PostgreSQL and Neon source connector. #15690
• Adds connector parameter ssl.mode for MySQL source connector. #16579
• Supports parameters snapshot.interval and snapshot.batch_size under WITH options when creating a table from a CDC source. #16426
• Supports implicitly converting numeric types from PostgreSQL sources into rw_int256 or varchar. #16346
• Supports configuring the timeout of CDC sources. #16598
• Supports timestamptz.handling.mode formatting option when creating a source with PLAIN, UPSERT, or DEBEZIUM JSON formats. #16265
• Only uses fragment ID as group ID for Kafka sources. #16111
• Supports cluster URLs for Redis sink connector. #16034
• Supports creating Delta sinks with GCS. #16182
• Supports Snowflake sink connector. #15429
• Supports creating upsert type BigQuery sinks. #15780

Installation and deployment

• Sets PostgreSQL as the default meta store when deploying with Docker Compose. #16724

Cluster configuration changes

• Supports using ALTER SYSTEM to set a system-wide default value for a session parameter. #16062
• Supports setting streaming_rate_limit to zero. #16333
• Supports configuring the reserved memory bytes of the compute node by using RW_RESERVED_MEMORY_BYTES runtime parameter and reserved-memory-bytes startup option. #16433
• Introduce new timeout and retry configurations for ObjectStore and deprecate ambiguous timeout configurations. #16231

Fixes

• Properly convert -inf, +inf, and nan types to null for JDBC sinks. #16230
• Handles sinking -inf, +inf, and nan types for ClickHouse, Doris, and StarRocks sink connectors. #15664
• Fixes an issue where DELETE events could not be sinked if the primary key is uuid type for JDBC sinks.#16447
• Fixes an issue where enum types from PostgreSQL could not be ingested as varchar types. #16423
• Fixes sources with encode avro on decimal ingesting. #16202
• Fixes sources with encode avro on bytes/fixed/decimal default value. #16414

v1.9.0

For installation and running instructions, see Get started.

Main changes

SQL features

• Query syntax:
  • Supports non-append-only temporal joins, where the outer side is not required to be append-only. #16286
• SQL commands:
  • Supports DISCARD ALL command for Npgsql compatibility. #16432
  • Supports creating, dropping, altering, and showing subscriptions. #14831
  • Supports cursors for subscription queues. #15180
  • Supports altering stream rate limit for sources and tables with a source. #16399
  • Supports RECOVER command to trigger an ad-hoc recovery. #16259
• SQL functions & operators:
  • Supports using a single ASCII character as an escape character with the LIKE clause. #16057
  • Supports jsonb_populate_record() and jsonb_populate_recordset(). #13421
• System catalog:
  • Supports system table rw_iceberg_files for displaying the files of an Iceberg source or table. #16180
  • Supports system table rw_iceberg_snapshot for listing all snapshots. #16175

Connectors

• Provides stable support for SQLAlchemy 2.0. risingwavelabs/sqlalchemy-risingwave#29
• Deprecates s3 connector. #16337
• Supports generated columns for non-shared CDC tables. #16522
• Supports time travel for Iceberg sources. #15866
• Blocks sink creation until backfill is completed by default. #16249
• Supports Kafka connector parameter properties.request.required.acks. #16482
• Adds connector parameter ssl.mode for PostgreSQL and Neon source connector. #15690
• Adds connector parameter ssl.mode for MySQL source connector. #16579
• Supports parameters snapshot.interval and snapshot.batch_size under WITH options when creating a table from a CDC source. #16426
• Supports implicitly converting numeric types from PostgreSQL sources into rw_int256 or varchar. #16346
• Supports configuring the timeout of CDC sources. #16598
• Supports timestamptz.handling.mode formatting option when creating a source with PLAIN, UPSERT, or DEBEZIUM JSON formats. #16265
• Only uses fragment ID as group ID for Kafka sources. #16111
• Supports cluster URLs for Redis sink connector. #16034
• Supports creating Delta sinks with GCS. #16182
• Supports Snowflake sink connector. #15429
• Supports creating upsert type BigQuery sinks. #15780

Installation and deployment

• Sets PostgreSQL as the default meta store when deploying with Docker Compose. #16724

Cluster configuration changes

• Supports using ALTER SYSTEM to set a system-wide default value for a session parameter. #16062
• Supports setting streaming_rate_limit to zero. #16333
• Supports configuring the reserved memory bytes of the compute node by using RW_RESERVED_MEMORY_BYTES runtime parameter and reserved-memory-bytes startup option. #16433
• Introduce new timeout and retry configurations for ObjectStore and deprecate ambiguous timeout configurations. #16231

Fixes

• Properly convert -inf, +inf, and nan types to null for JDBC sinks. #16230
• Handles sinking -inf, +inf, and nan types for ClickHouse, Doris, and StarRocks sink connectors. #15664
• Fixes an issue where DELETE events could not be sinked if the primary key is uuid type for JDBC sinks.#16447
• Fixes an issue where enum types from PostgreSQL could not be ingested as varchar types. #16423
• Fixes sources with encode avro on decimal ingesting. #16202
• Fixes sources with encode avro on bytes/fixed/decimal default value. #16414

Full Changelog: v1.8.0...v1.9.0

v1.8.2

release v1.8.2

v1.8.1

release v1.8.1

v1.8.0

For installation and running instructions, see Get started.

Main changes

SQL features

• Query syntax:
  • Supports RANGE frames in window function calls. #14416
• SQL commands:
  • Adds support for embedded Rust UDFs. #14903
  • Adds support for embedded Python UDFs. #15168
  • Supports refreshing the schema of a table created using an external connection to get the latest schema. #15025
  • Supports refreshing the schema of a source to get the latest schema. #15541
  • Adds a description column to the system parameters table. #15113
  • Supports authenticating with OAuth token acquired from the Cloud when creating a user. #13151
• SQL functions & operators:
  • Supports ruby-pg. #14859
  • Supports VARIADIC arguments for the functions format, concat_ws, jsonb_build_array, jsonb_build_object, jsonb_extract_path, jsonb_extract_path_text. #14753
  • Supports concat function. #14753
• System catalog:
  • Adds missing columns for  pg_catalog.pg_index, rw_catalog.rw_columns and information_schema.columns, and system view pg_catalog.pg_partitioned_table. #15151
  • Supports pg_catalog.pg_constraint for DBeaver compatibility. #15227
  • Supports pg_catalog.pg_stat_get_numscans for DBeaver. #15642
  • Supports system table rw_depend. #15385
  • Supports pg_settings catalog. #15108

Connectors

• Breaking change: Sinks created from v1.6 and earlier that have decouple enabled may cause compatibility issues. Check if you have any sinks with this configuration by using the internal table rw_sink_decouple before upgrading to v1.8. #15613
• Avro tables and sources now require a schema registry during creation. #15256
• Supports using Karapace when specifying a schema registry. #15486
• Supports Protobuf data format for NATS JetStream source. #15378
• Supports FORMAT PLAIN ENCODE BYTES for NATS source connector. #15806
• Supports Confluent schema registry for Kafka sinks when using FORMAT PLAIN ENCODE PROTOBUF. #15546
• Adds Kafka sink and source parameter enable.ssl.certificate.verification. #15073
• Supports max_batch_rows and  request_timeout parameters for Cassandra and ScyllaDB sources. #15516
• Adds built-in MongoDB CDC source connector. #14966
• Adds ignore_option parameter for sources created using Debezium format. #15304
• Supports batch read from Iceberg source. #15214
• Supports automatically deriving columns from Iceberg source. #15415
• Supports JDBC catalog for Iceberg sources. #15551
• Adds JDBC and Hive catalogs for Iceberg sink. #14885

Installation and deployment

Supports tab-completion for SET and ALTER SYSTEM SET commands in psql client. feat(frontend): support tab-completion for [ALTER SYSTEM] SET

Full Changelog: v1.7.0...v1.8.0

v1.7.3

release v1.7.3

v1.7.2

release v1.7.2

v1.7.1

release v1.7.1

v1.7.0

For installation and running instructions, see Get started.

Main changes

SQL features

• Query syntax:
  • Supports using * when creating a table or source with schema from an external connector to read all columns. #14644
  • Supports INCLUDE clause to add additional connector columns. #14215
  • Supports INCLUDE HEADER clause to specify desired keys in header and make it a column. #14628
• SQL commands:
  • Supports changing the schema registry by redefining the format_encode_options. #14057
• SQL functions & operators:
  • Adds experimental support for JavaScript UDF. #14513
  • Adds experimental support for Rust UDF. #14271
  • Supports implicit type cast for UDF. #14458
  • Supports named SQL UDF. #14806
  • Supports encrypt and decrypt functions. #14717
  • Supports make_date(), make_time(), and make_timestamp(). #14827
• System catalog:
  • Change rw_streaming_parallelism to allow queries on streaming job parallelism with job name and type. Adds system view rw_fragment_parallelism to allow for queries on parallelism information at fragment level #14789, #14261
  • Adds relpersistence in pg_class catalog. #14400
  • Supports pg_get_viewdef() . #14336

Connectors

• Cassandra and ScyllaDB sinks no longer support timestamp type. #14413
• Updates StarRocks sink connector parameters. #14823
• Introduces snapshot option to allow users to disable CDC backfill and to only consume from the latest changelog. #14718
• Sets the default value of transactional parameter to true for MySQL and Postgres CDC shared sources. #14899

Installation and deployment

[Pre-Release] Supports single_node node to run RisingWave in a single process.

Cluster configuration changes

• Introduce a session variable batch_enable_distributed_dml to enable batch ingesting. #14630
• Changes wording from AUTO to ADAPTIVE parallelism. #14414
• Supports adaptive scaling for streaming jobs by default. #14873

Full Changelog: v1.6.0...v1.7.0

v1.7.0-standalone

release v1.7.0-standalone