Github user hbdeshmukh commented on the issue:
https://github.com/apache/incubator-quickstep/pull/281
Hi @zuyu
The changes look good to me. Can you please try the following scenarios to
make sure the scheduling works fine? These are some of the corner cases that I
can think of:
1. Join of two empty tables.
2. Join of two empty tables with scan on one table.
3. Join of two empty tables with scan on both tables.
4. One empty table joined with another non-empty table.
5. One empty table joined with another non-empty table with a scan.
6. One empty table with scan joined with another non-empty table with a
scan.
7. Any variation from above with an aggregation.
8. A non-equi join.
---
