Fork me on GitHub
#xtdb
<
2021-01-01
>
Steven Deobald00:01:12

Might be less confusing for folks who are new to Clojure if they’re commented out anyway? :woman-shrugging:

roelof00:01:03

and is crux now a im memory database or does really data get stored on my disk

gklijs01:01:09

By default it's persisted on Kafka I think. But other options are available.

refset02:01:15

Actually nothing gets persisted by default. Everything is in-memory by default. You have to specify all the configuration (RocksDB, Kafka or whatever) explicitly in the call to start-node and include the relevant modules via your classpath deps

roelof11:01:41

oke, thanks, then now I understand why data is lost when I start up repl again

refset00:01:13

🙂 cool & sorry that's not clearer from the docs - I'll give it some thought!

Toyam Cox07:01:30

I hate to be That Guy, but why :in replacing :args? I understood the semantics of :args at this point, I'm still trying to wrap my head around :in

refset00:01:23

I sympathise 🙂 :in was certainly more complex to grasp at first, but :args is much less powerful and keeping both approaches around for the long-term would only complicate things further. Specifically, the benefit is that the inputs provided via :in are handled as first-class relations, whereas to achieve the same effect with :args the user has to generate the cartesian products of all possible input combinations before handing them to the query engine (which could be prohibitively large for certain kinds of queries). In many scenarios :in will avoid a lot of work for the user and for the query engine compared to the :args alternative

Toyam Cox03:01:47

I understand, having written some of that code myself. Which just now needs to be changed 🙂

Toyam Cox07:01:27

Also is it no longer idiomatic to put question-marks before query bindings?

dominicm09:01:14

That's a datomic idiom. Crux has never followed it really. The query engine doesn't care about any prefixes.