Improve DuckDB presence within IntelliJ IDEA

[rsmckinney]

As the screen capture shows, DuckDB is available as a data source within IntelliJ IDEA. However, DuckDB is in the basic support category, as opposed to the "complete support" category.

This is a real disappointment when using DuckDB in IntelliJ in terms of SQL rendering, code completion, etc.

Even with the dialect set to PostgreSQL, queries that use extensions fail to parse in the editor. Usage of all those awesome extensions render in red and fail code completion.

Would be nice to work with JetBrains to level up here, it would help DuckDB gain traction with Java.

[fourcadefran]

And Webstorm ❤️

[bsplosion]

Adding a plus one here as well. I've been (sort of) getting by with connections leveraging the Postgres dialect, but I'm hitting very frequent situations where the development environment is flagging queries as broken when they're just using DuckDB dialect in reality. Connections work fine, and running the "broken" queries will work if they're actually valid.

We know about DBeaver, but we already live in the Jetbrains IDEs for literally everything else, so this would be extremely helpful. We don't use tools like pgAdmin, SSMS, etc today for other database interactions since it doesn't make sense to have our developers in multiple places, but DuckDB seems like... the odd duck out [rimshot].

[dillongardner]

Completely agree. My primary database is Postgres and heavily use DuckDB with the postgres extension to parse external parquet files before writing them to my database. The lack of support for DuckDB makes this painful.

A testament to the quality of what DuckDB offers that the pain is worth it.

[bsplosion]

With the launch of Quack protocol (https://duckdb.org/quack/) is there interest in re-engaging this ticket? We've ended up leveraging Duckgres (https://github.com/PostHog/duckgres) plus a monumental amount of monkeypatching to get a fully Postgres-compliant remote service up and running for our developers to use for operations, but that's a hack and likely very brittle over time. Getting first-class supported plugins out into the ecosystem would be really helpful.

[staticlibs]

@bsplosion

Yes, we definitely would like to support more tools that use JDBC driver. Just the gap between "basic things work, data can be shown" and "tool is fully supported" can be pretty large. And every such tool needs to be addressed separately (example: #573).

In general, reworking the JDBC API data fetching on top of DuckDBDataChunkReader (#682) and improving support for extended types (LIST, STRUCT etc) in JDBC API/Appender/UDFs have higher priority.