Georg Traar
Georg Traar
> @proddata Would this still work at 100,000 rows? I suppose. Quick test with 1 mio rows of 10 values takes a few seconds already. With 100.000 it is still...
Hi @bputt-e Just to be sure, are you really using 4.6.4 (in testing channels) or 4.6.3? That issue originally occurred with 4.6.x, but should have been fixed with 4.6.4. However...
This probably could also be relevant for queries like "first/last" queries mentioned in #10528 ```sql SELECT device, speed FROM data a, ( SELECT max(ts) AS ts, device FROM data GROUP...
I think it would make sense to maybe mention them.
Another approach: ```sql CREATE OR REPLACE FUNCTION pos_max(ts ARRAY(TIMESTAMP)) RETURNS long LANGUAGE JAVASCRIPT AS 'function pos_max(ts) { return ts.reduce((iMax, x, i, arr) => x > arr[iMax] ? i : iMax,...
**CTE + JOIN** ```sql WITH cte AS (SELECT device_id, MAX(ts) ts FROM data GROUP BY 1) SELECT r.* FROM data r JOIN cte using(device_id,ts) WHERE r.time >= (SELECT min(ts) FROM...
[Timescale](https://docs.timescale.com/api/latest/hyperfunctions/last/), [SingleStore](https://docs.singlestore.com/managed-service/en/reference/sql-reference/time-series-functions/first.html), [Elasticsearch](https://www.elastic.co/guide/en/elasticsearch/reference/current/sql-functions-aggs.html#sql-functions-aggs-first) use `FIRST/LAST` Presto uses `max_by / min_by` Clickhouse uses `argMax / argMin` Druid uses `LASTWITHTIME / FIRSTWITHTIME` Influx uses `last()/first()` I don't have a super strong preference...
@mfussenegger It is from the PostgreSQL docs directly ... https://www.postgresql.org/docs/16/sql-createrole.html > The CREATE ROLE statement is in the SQL standard, but the standard only requires the syntax > CREATE ROLE...
The adjustments to `crate` as default catalog have been largely done. For the different behaviour regarding the `search_path` I created another issue: https://github.com/crate/crate/issues/13472
@matriv Am I understanding it correctly, that the `_` underscore part is missing? If so, I think it might make more sense to create a separate issue.