JetBrain's DataGrip and IDE connectors

[amotl]

DataGrip provides a summary which information is needed to come up with a dedicated connector: https://www.jetbrains.com/datagrip/new_dbms/

Originally posted by @ckurze in #35 (comment)

[amotl]

It looks like JetBrains may have depublished their page? Here is an archived variant, but maybe you can find back a native source again?

https://web.archive.org/web/20240129181954/https://www.jetbrains.com/datagrip/new_dbms/

/cc @surister, @ckurze

[surister]

I'm joining https://plugins.jetbrains.com/slack/ I will ask there :), I'd love to get CrateDB support for jetbrains, such a must for developer experience IMHO.

[surister]

Asked again in their slack https://jetbrains-platform.slack.com/archives/C5P9YB0LT/p1733319820077659

[amotl]

Hi. We've refreshed the documentation about how to connect to CrateDB using DataGrip and the CrateDB JDBC Driver.
Currently, the driver needs to be supplied manually by downloading and providing a JAR file.

https://cratedb.com/docs/guide/integrate/datagrip/

References

• [Meta] DBeaver / Datagrip - PG connector compatibility tracker  crate#12101 (comment)
• Jetbrains IDE DataGrip compatibility crate#15062 (comment)
• DataGrip Limited Compatibility

[amotl]

Over here, we have a bug that needs to be addressed, most probably by adding a corresponding behavioral/dialect adjustment to the CrateDB JDBC Driver, when possible.

• DataGrip fails to create tables with uppercase names crate-jdbc#441

If not, we really need to emphasize into getting a dedicated adapter for CrateDB into JetBrain's registry, where @surister is already looking into.

/cc @kneth, @zolbatar

[surister]

Jetbrain's devrel told me to follow up on https://youtrack.jetbrains.com/issue/DBE-12075/CrateDB-Full-Dialect-Support, left a comment there but it seems that it's up to them. I'll try to be active and follow up swiftly with them.

[amotl]

Thank you very much, @surister. So, the gist is:

The better your jdbc driver is implemented, the better is our basic support.

Indeed, CrateDB JDBC Internals outlines quite a bunch of things that can be done on the driver level already, essentially all the ingredients of that delta, plus something in the io.crate.client.jdbc wrapper. This is where @kneth is currently looking into. 🙇

What I almost certainly know is the detail that this gap of CrateDB, for example, seems to be masked, or otherwise not be an issue with the CrateDB JDBC driver, while it is an issue with current vanilla pgJDBC.

• Support pg_catalog.txid_current() crate#12086

So, effectively, the CrateDB JDBC Driver should aim to compensate for all possible details like outlined per https://github.com/crate/crate/labels/tool%3A%20DataGrip, or otherwise CrateDB may improve on some of them? ¹

NB

Within SQLAlchemy Core, we are doing many such compensations and masking in the act of re-using as many details from the PostgreSQL dialect adapter as possible, and adjusting them properly for the SQLAlchemy CrateDB Dialect. From our experiences there, we think it is business as usual, and should likewise be applied to Java/JDBC and downstream frameworks. It's too sad that DataGrip is closed source.

Footnotes

1. However, on a few significant on them, that might not be easy, and will take time. ↩