Clarification requested on KUDU.
In the KUDU guides the following is stated:
Row delete and update operations must also specify the full primary key of the row to be changed. Kudu does not natively support range deletes or updates.
The first part makes sense. However, using IMPALA via Hue I can easily issue commands like these that relate to the highlighted part of the prose:
delete from metrics_001 where (value >= 400 and value <= 600);
update metrics_001 set value = value + 1000 where (value >= 600 and value <= 800);
which execute as expected.
Does the statement mean that IMPALA allows this to be so? Could not find it from the documentation. I must be missing something elementary.
Impala first scans Kudu for the records that meet the filter criteria, and then it sends back to Kudu the individual delete/update operations for each key that was found.