Jeremy Taylor
Jeremy Taylor
e.g. `:xtdb.api/delete` should be conformed with the original `:xt/id` if it is still available, otherwise there are very limited options for reversing the lookup of an `#xtdb/id` (without resorting to...
Based on a report from @jacobobryant via https://discuss.xtdb.com/t/where-clause-not-working/34 This should never cause an OOM: ```clojure (with-open [l (xt/open-tx-log (xtdb-node) nil true)] (while (.hasNext l) (.next l))) ``` This is using...
Discussion: https://clojurians.slack.com/archives/CG3AM2F7V/p1651517243301129
Kudos to @rhishikeshj for figuring this out and reporting it. His specific issue was caused by attempting to re-use an existing Kafka topic where he had previously manually published a...
This isn't a severe issue since MemKv is not generally intended for production usage, however it has come up a couple of times in the various forums, e.g. https://juxt-oss.zulipchat.com/#narrow/stream/194466-crux-users/topic/Checkpointer.20restoring.20failed
We don't necessarily need to support this, but it was unexpected. ```clojure (-> (crux/db (crux/start-node {})) (crux/with-tx [[:crux.tx/put {:crux.db/id 1 :a 2}]]) (crux/with-tx [[:crux.tx/put {:crux.db/id 2 :a 2}]])) ``` ```...
e.g. ```clojure (or [e :att "val1"] [e :att "val2"]) ``` ...could be re-written as... ```clojure [e :att #{"val1" "val2"}] ``` The difference between them is that the `or` version will...
All bound `:in` arguments should be included early in the join order to improve performance. Failing slowdown test: https://github.com/refset/crux/commit/617079c58d368de1c667bbc9d4684abc3bd83fef#diff-6c6576892a723f0244c94ef6f4ce3132aa1d8e6280f07a59e2d90c7580eef6a4R3689-R3719
String IDs weren't supported back when the SPARQL implementation was written, so this conditional worked pretty effectively: https://github.com/xtdb/xtdb/blame/e2f51ed99fc2716faa8ad254c0b18166c937b134/labs/rdf/src/xtdb/rdf.clj#L237 However, since strings are now valid IDs, that `if` can't be relied...
There's been a report of the node re-indexing process silently halting at regular intervals (e.g. approx. every N thousand transactions). It can be worked around by restarting the node repeatedly,...