Fork me on GitHub
#crux
<
2020-04-18
>
bocaj15:04:18

Crux seems ideal for a traditional data warehouse. Do you know of a use case? What downsides would I need to address?

bocaj16:04:53

One limit I see using the jdbc node, is the number of rows in the tx_events table. For mssql, this is limited by storage

taylor.jeremydavid11:04:03

I've not yet spoken to anyone actively using Crux as a data warehouse as such, but certainly I see a lot of overlaps in capabilities. The biggest limitation is probably the shear amount of data that people want to store in a DW precludes Crux from being practical. Crux might be more suitable as a metadata system for a DW.

taylor.jeremydavid11:04:16

In regards the row limit, I guess we could shard across multiple tables once they're "full" without too much issue.

bocaj17:04:47

Will crux support SQL Server in the future, or can I override configuration to support mssql until it's available?

richiardiandrea23:04:10

I think I saw in the previous Crux demo that it was already implemented in alpha on top of Calcite

taylor.jeremydavid11:04:56

Our Calcite work is to support SQL querying (via JDBC) but this is unrelated to crux-jdbc, which is exclusively about the underlying storage of transactions and documents in JDBC stores. Adding SQL Server support for crux-jdbc is definitely possible and something we would like to facilitate. @ if you send us a brief email to <mailto:[email protected]|[email protected]> I'll add you to our beta Zulip stream to discuss possible timelines. In the meantime, forking your own crux-jdbc module to add mssql support is quite a feasible task though. There shouldn't be many changes required and there are plenty of tests to help sanity check 🙂 (and PRs are welcome too!)

taylor.jeremydavid15:04:46

@ SQL Server support is in master 🙂 https://github.com/juxt/crux/pull/811 /cc @ @

richiardiandrea15:04:42

Nice and sorry for barging in 😄