DROP should have a flag to terminate dependent queries

[rmoff]

In v5.0.0-SNAPSHOT, new behaviour when trying to drop a table that is being populated by a query:

ksql> CREATE TABLE PURCHASE_COUNT_BY_CITY_PER_30MIN AS SELECT city,COUNT(*) FROM Purchase_enriched WINDOW TUMBLING (SIZE 30 MINUTES) GROUP BY city;

 Message
---------------------------
 Table created and running
---------------------------

ksql> DROP TABLE PURCHASE_COUNT_BY_CITY_PER_30MIN;
Cannot drop the data source. The following queries read from this source: [] and the following queries write into this source: [CTAS_PURCHASE_COUNT_BY_CITY_PER_30MIN]. You need to terminate them before dropping this source.
ksql>

I understand the need for this protection, but for UX it would be really useful if we can have a flag to automatically do the required termination so that the object can be dropped. The common usecase will be a mistake in the DDL requiring a drop/recreate.

I would suggest something like DROP TABLE FOO CASCADE;

[rmoff]

To be clear, the reason for adding something like CASCADE is to make the development experience nicer to users. If one is developing code and iterating on ideas, dropping & recreating objects is going to be common, and having to manually terminate the query each time literally doubles the effort.

[DW-Pete]

It is also a common use case to promote code using scripts and something like Jenkins; typically we may want to:

drop abc if exists;
create abc;

so that the topics, streams and tables being revised in a deployment are replaced cleanly. Active queries stop the run script and terminate in a KSQL script is not an option

[rmoff]

Closing to track in #2177