James Henderson
James Henderson
re #41 - we might as well include ByteBuffer support too?
Hey again @benedekfazekas :smile: Looks like `mranderson` is replacing imports that it shouldn't when it needs to rename only _some_ classes within a package. repro: * `^:inline-dep [potemkin "0.4.5"]` -...
Hey again :smile: [clj-tuple](https://github.com/ztellman/clj-tuple) (a transient dependency of `clj-http`, in my case) has a top-level `clj-tuple` namespace, which isn't getting included in `:import` replacements. I reckon it's because `clj-files->dirs` has...
Workaround: ensure you've [configured a persistent local document-store](https://docs.xtdb.com/storage/1.21.0/kafka/) on your Kafka doc-store. (and rebuild your index-store, if needed) Kafka's document store uses a local document store for fast random access,...
Currently, when required to reconstruct documents (e.g. for `entity` or `pull [*]`) we have to go to the document store, because the query indices don't contain enough information. Specifically, they...
see #1514 - thanks @philippkueng :pray: ```clojure (time (count (crux/q (crux/db node) '{:find [(pull ?e [:title :date ...]) ?date] :where [[?e :date ?date]] :order-by [[?date :desc]] :limit 10}))) ``` The...
We've now had a few issues come in regarding our use of content hashing in Crux. We (essentially) use `(sha-1 (->nippy v))` to calculate a content hash. This has caused...
Raised by @brdloush on Slack - thanks to both you and @mvarela for raising and helping to track it down :pray: Fetching entities by id remains fast; fetching entities where...
(from @refset) query parameter values must not have any non-Serializable metadata attached, otherwise you get this weird error `xtdb.IllegalArgumentException: invalid tx-op: Unfreezable type: class ...` * add check of some...
One of the first questions we often ask users with query performance issues is to get us the debug logging for `crux.query`, which contains the attribute stats and join order...