I came across the 'Reading RisingWave from Spark or Presto' project listed on the RisingWave code camp projects page. I'm curious to know if this feature is currently blocked due to lack of resources or if there's a planned timeframe for its implementation. This feature is particularly important for us as we already have Presto set up and use it as our query federation layer. It would be beneficial to join data in RisingWave with other data sources and leverage Presto's capabilities for running analytical SQL queries efficiently.
Nizar Hejazi
Asked on May 11, 2023
The feature to read RisingWave from Spark or Presto is still experimental and was more complicated than initially expected. A demo with minimum functionality has been implemented to query RisingWave materialized view data with Apache Spark and Trino. Presto support should be simple to add due to its similar plugin framework to Trino's. Currently, RisingWave can be queried via JDBC as it is compatible with PostgreSQL, and Presto has JDBC support. More performance benchmarks and experiments will be conducted to compare the performance with other systems like Apache Iceberg. The feature will be contributed to the Trino and Presto repositories once it is mature enough, and its development can be prioritized based on user requests.
