fix(batch scheduler): has_table_scan had false negative

[xiangjinwu]

What's changed and what's your intention?

Found this issue in #3172, where stage 4 and 7 were marked as has_table_scan = false with Fragment:

BatchProject { exprs: [$0, ($1 * (1:Int32 - $2))] }
  BatchFilter { predicate: ($3 >= '1993-01-01':Varchar::Date) AND ($3 < ('1993-01-01':Varchar::Date + '3 mons 00:00:00':Interval)) }
    BatchScan { table: lineitem, columns: [l_suppkey, l_extendedprice, l_discount, l_shipdate] }

Hard to test its e2e impact.

Checklist

• I have written necessary docs and comments
• I have added necessary unit tests and integration tests
• All checks passed in ./risedev check (or alias, ./risedev c)

Refer to a related PR or issue link (optional)

[liurenjie1024]

Generally LGTM, please add some tests.

[liurenjie1024]

Not sure what the actual impact is. No tests were added.

It may early release hummock snapshot.

[codecov]

Codecov Report

Merging #3175 (b77fb67) into main (6d98333) will increase coverage by 0.00%.
The diff coverage is 90.00%.

@@           Coverage Diff           @@
##             main    #3175   +/-   ##
=======================================
  Coverage   73.57%   73.57%           
=======================================
  Files         744      744           
  Lines      102180   102191   +11     
=======================================
+ Hits        75174    75192   +18     
+ Misses      27006    26999    -7     

Flag	Coverage Δ
rust	73.57% <90.00%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/frontend/src/scheduler/distributed/query.rs	67.20% <50.00%> (ø)
src/frontend/src/scheduler/plan_fragmenter.rs	93.64% <100.00%> (+0.42%)	⬆️
src/meta/src/hummock/mock_hummock_meta_client.rs	42.39% <0.00%> (-1.09%)	⬇️
src/storage/src/hummock/local_version_manager.rs	84.12% <0.00%> (-0.16%)	⬇️
src/common/src/types/mod.rs	68.07% <0.00%> (+0.16%)	⬆️
...c/frontend/src/optimizer/plan_node/batch_filter.rs	97.43% <0.00%> (+17.94%)	⬆️

📣 Codecov can now indicate which changes are the most critical in Pull Requests. Learn more

[xiangjinwu]

Not sure what the actual impact is. No tests were added.

It may early release hummock snapshot.

So it is hard to test e2e and we shall just leverage unit tests. Right?

[liurenjie1024]

Yes, it hard to test in e2e.

[lmatz]

because our plan is a DAG now?

I see.... It's stage level instead of node level flag.

[xiangjinwu]

because our plan is a DAG now?

No. Because the boolean flag is set back to false when the recursion returns thru ancestors 😂