For installation and running instructions, see Get started.

Main changes

Installation

• Now, you can easily install RisingWave on your local machine with Homebrew by running  brew install risingwave.

Administration

• Adds the pg_indexes and dattablespace system catalogs. #9844, #9822
• Now, the SHOW PARAMETERS command will display the mutability of each system parameter. #9526

SQL features

• Experimental features: Adds support for 256-bit integer. #9146, #9184, #9186, #9191, #9217
• Indexes can be created on expressions. #9142
• Adds support for expressions in aggregate function arguments. #9955
• Adds support for VALUES clause. #8751
• Adds support for generated columns, which are special columns computed from other columns in a table or source. #8700, #9580
• Adds support for using expressions in the ORDER BY and PARTITION BY clauses. #9827
• New SQL commands
  • CREATE CONNECTION and SHOW CONNECTIONS: Creates an AWS PrivateLink connection and show all available connections. #8907
  • DROP CONNECTION: Drops a connection. #9128
  • SHOW FUNCTIONS: Shows existing user-defined functions. #9398
  • DROP FUNCTIONS: Drops a user-defined function. #9561
  • SHOW CREATE SOURCE and SHOW CREATE SINK: Shows the SQL statement used to create a source or sink. #9083
  • SHOW INDEXES: Shows all indexes on a particular table. #9835
• SQL functions
  • Adds support for trigonometric functions. #8838, #8918, #9064, #9203, #9259
  • Adds support for degrees and radians functions. #9108
  • Adds support for the lag() and lead() window functions and the OVER() and  EMIT ON WINDOW CLOSE clause. #9597, #9622, #9701
  • Adds support for new aggregate functions, including bool_and, bool_or, jsonb_agg, and jsonb_object_agg. #9452
  • Adds support for max(), min(), and count() for timestamptz data. #9165
  • Adds support for microseconds and milliseconds for to_char() and to_timestamp(). #9257
  • Adds support for multibyte Unicode in overlay() and ascii() functions. #9321
  • Adds support for the string_to_array() function. #9289
  • Adds support for array_positions(). #9152
  • Adds support for cardinality(). #8867
  • Adds support for array_remove(). #9116
  • Adds support for trim_array(). #9265
  • Adds support for array range access. #9362
  • Adds support for JSONB in UDF. #9103
  • Adds support for btrim() and updates trim() to PostgreSQL standard syntax. #8985
  • Adds support for date_part(). #8830
  • Expands extract() with more fields. #8830
  • Adds support for proctime(), which returns the system time with time zone when a record is processed. #9088
  • Adds support for translate(), @(), and ceiling(). #8998
  • Adds support for encode() and decode(). #9351
  • Adds support for the intersect operator. #9573
  • Adds support for the default escape \ in a like expression. #9624
  • Adds support for the IS [NOT] UNKNOWN comparison predicate. #9965
  • Adds support for the starts_with() string function and ^@. #9967
  • Adds support for unary trunc, ln, log10 (log), exp, cbrt (||/) mathematical functions. #9991

Connectors

• Adds support for ingesting CDC data from TiDB and sinking data to TiDB with the JDBC connector. #8708
• Adds support for ingesting CDC data from Citus. #8988
• Adds support for loading Pulsar secret key file from AWS S3. #8428, #8222
• Adds support for using an established AWS PrivateLink connection in a CREATE SOURCE, CREATE TABLE, or CREATE SINK statement for a Kafka source/sink. #9119, #9128, #9728, #9263
• Deprecates the use_transaction field in the Kafka sink connector. #9207
• Adds support for zstd compression type for Kafka connector. #9297
• Deprecates the upsert property in the Kafka connector as it can be inferred from the row format. #9457
• Adds a new field properties.sync.call.timeout in the WITH clause of the Kafka source connector to control the timeout. #9005
• Adds support for a new row format DEBEZIUM_MONGO_JSON in the Kafka source connector. #9250
• Adds CSV format support for the Kafka source connector. #9875

Cluster configuration changes

• --data_directoryand --state_storemust be specified on CLI of the meta node, or the cluster will fail to start. #9170
• Clusters will refuse to start if the specified object store URL identified by state_store and data_directory is occupied by another instance. Do not share the object store URL between multiple clusters. #9642

Full Changelog: v0.18.0...v0.19.0