lloyd tabb
lloyd tabb
Nice thought! I looked at this but it requires two reads of 'movies' in this case, which I think is unreasonable. You could do this faster by `SELECT MAX(array_len(movies.genres)) from...
[Malloy](https://github.com/looker-open-source/malloy/), our data language, is generating this code, so I need a pretty generic solution. The problem with taking a pass over the data is that I'd have to push...
> WHERE base.names[names_0.__row_id]."percent_in_state">1 If you remove the above where clause, you should see results where `percent_in_state` is a value between 0 and 1. There should be no way for it...
So, IMO, pivoting is a rendering thing. The real issue is limiting the dataset so that you can fill a grid appropriately. When pivoting two high cardinality dimensions (or sets...
It only seems to fail on DuckDB :(
Looks like the reason is that were are getting computation in this column from nested datasets. A possible solution is to bracket with a case ``` CASE WHEN group_set =...
For debugging ``` -- connection: duckdb WITH __stage0 AS ( SELECT group_set, base."faa_region" as "faa_region__0", CASE WHEN group_set=0 THEN COUNT(1) END as "airport_count__0", ROW_NUMBER() OVER(PARTITION BY group_set ORDER BY CASE...
@mtoy-googly-moogly for discussion.
Yes, for now. I think it is too confusing to know when it might and might not work. We'll end up creating duplicate records in many cases. I think forcing...