[CALCITE-3916] Implement top-down rule applying and upper bound space pruning #1950
Conversation
| CheckInput(OptimizeInputs context, | ||
| RelNode parent, CascadeRelSubset input, int i, RelOptCost upper) { | ||
| this.context = context; | ||
| this.parent = parent; |
There was a problem hiding this comment.
Is there any test for it?
|
|
||
| /** | ||
| * An enum representing optimizing state of current RelSubset | ||
| */ |
There was a problem hiding this comment.
What's the difference between CascadeRelSubset and RelSubset?
There was a problem hiding this comment.
CascadeRelSubset have two more fields: optimizeState and upperBound. The comments in this class will help you understand what they are used for.
|
hi, @FatLittle Thank you very much for the PR you submitted. Can you include the issue link in the description of this PR? |
|
@XuQianJin-Stars I think this is the most related JIRA: |
That's it. Thanks, Haisheng! |
FatLittle
changed the title
FatLittle
force-pushed
the
odps-cascade
branch
2 times, most recently
from
1126e7b
to
4d6e477
Compare
| protected boolean cascade() { | ||
| return false; | ||
| } | ||
|
|
pruning in calcite
pruning in calcite: add more tests and resolve some comments
FatLittle
force-pushed
the
odps-cascade
branch
4 times, most recently
from
9cfa69e
to
54cde8d
Compare
pruning in calcite: fix case failure
|
added a commit. Propagate Traits with Haisheng's trait propagation frameworks. Remove AC dependencies. |
| * CascadePlanner that apply memo pruning logic` | ||
| * based on the data structure of VolcanoPlanner | ||
| */ | ||
| public class CascadePlanner extends VolcanoPlanner { |
There was a problem hiding this comment.
I think the new planner's name should be CascadesPlanner not CascadePlanner according to the paper "The Cascades Framework for Query Optimization", and we can also find “Cascades” in the Columbia planner source code.
There was a problem hiding this comment.
I have modified the names in my new PR: #1991
| } | ||
|
|
||
| RelOptCost lowerBound = null; | ||
| for (RelNode relNode : subset.getRels()) { |
There was a problem hiding this comment.
If current group is not fully optimized and has no winner, there may be problems to get the lower bound here. Since the current group's optimization may affact the input groups of the current group. E.g., the current group's optimization may push down some require traits and maybe somewhere of the inputs subtrees may enhance there's costs. So if we get the lower bound here, it may be larger than the actual lower bound which may miss optimal plan
|
This PR is superseded by #1991. |
Link to https://issues.apache.org/jira/browse/CALCITE-3916