You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The legacy search could answer a computation from a cheaper related one. The planning stages don't do this yet. The results are correct; only cheaper plans are missing.
Rollup: answer a coarse group by from a finer one (rollup).
Top-k reuse: answer a small-k top-k from a larger-k one over the same partition (topk_reuse).
Merge: collapse sum by (…) (sum_over_time(…)) into one accumulator.
These belong to Pass 2 (ASAP-aware CSE) as sharing rules, each producing a shared candidate that plan selection prices against the separate ones. The legacy implementations are at e80cfef (rollup.rs, topk_reuse.rs, and the nested-sum handling in replacement.rs) for reference.
Found in the #676 capability audit. Related: #580 and Pass 2.
Activity
feat(planner): nested sum merge as a Pass 1 candidate (#680)
feat(planner): Pass 2 rollup and exact top-k reuse rules (#680)
test(planner): reuse candidates selected and executed end to end (#680)
fix(planner): nested sum merge never groups by the dropped metric nam…
fix(planner): rollup only over Float64 values; cap rollup and top-k s…
test(planner): reuse candidates execute to the rows they replace (#680)
feat(planner): Pass 2 reuse rules (rollup, exact top-k reuse) and nes…