OFFSET queries are painfullly slow / do not complete

[okennedy]

mimir> LOAD 'https://raw.githubusercontent.com/thebuffalonews/shootings/master/data_cleanup/old_data/shootings_geo_dirty.csv' into shootings with staging;

The following query is super fast

mimir> select * from shootings limit 1;

Warm (after loading): 323 ms Cold (needs Spark init): 4891 ms

The following query takes > 60s to complete (I have yet to see an offset query finish)

mimir> select * from shootings limit 1 offset 1;

The explain plan is fine Mimir-side, this might be an issue with translating into Spark. This is a recent bug, since tabbing through the Vizier spreadsheet worked fine a month ago.

[okennedy]

Possibly already fixed by @mrb24