fix(tesseract): Resolve join for hint-less member-expression measures

[ovr]

A dependency-free member-expression measure such as count(*), in a query with no dimensions or filters to seed the join, resolved an empty join-hint set. That path calls JoinGraph.buildJoin([]), which returns null, and the Rust bridge then fails to deserialize the missing JoinDefinitionStatic ("invalid type: unit value, expected struct JoinDefinitionStatic").

In the multi-fact join grouping, fall back to the measure's own cube when its hint set is empty — but only for real, joinable cubes. View members are left untouched: their join comes from the other query members, and the view itself is not a joinable cube.

[claude]

Claude finished @ovr's task in 44s —— View job

---

Code Review

Verdict: LGTM — focused, well-tested fix. 0 high, 0 medium, 2 low.

• Read changed files
• Analyze fix logic in multi_fact_join_groups.rs
• Review test coverage
• Check edge cases (views, missing cubes, errors)

Summary

The fix in multi_fact_join_groups.rs:218-222 correctly addresses the root cause: when a measure-expression's hints are empty AND the query has no dimensions/filters to seed the join, falling back to the measure's own cube avoids the null-join path that the bridge can't deserialize. Guarding the fallback by is_view is the right call — views aren't joinable cubes and their join is resolved through other query members.

The new integration test (test_expr_measure_count_star_no_hints) is a clean regression test that exercises exactly the scenario described in the PR body, and the snapshot is minimal.

Low-severity notes

1. .ok() silently swallows lookup errors (multi_fact_join_groups.rs:254-259)

let is_view = query_tools
    .cube_evaluator()
    .cube_from_path(cube_name.clone())
    .ok()
    .and_then(|cube| cube.static_data().is_view)
    .unwrap_or(false);

If cube_from_path fails for any reason other than "cube doesn't exist" (e.g., a future evaluator error), it silently defaults to is_view = false, which then injects a JoinHintItem::Single(<bad name>) and the next join_for_hints call surfaces a less obvious error than the original lookup. Not a blocker since the current cube_from_path only fails on missing cubes, but propagating the error with ? would be a touch more defensive.

2. View detection happens per-measure (multi_fact_join_groups.rs:218-222)

Each empty-hint measure triggers an independent cube_from_path call. In practice multi-measure queries with all-empty hints are vanishingly rare so this is fine, but if it ever becomes hot it can be memoized on State.

Nits

• The doc comment block on lines 211–217 is excellent context — keep that style.
• fallback_hints_for_measure is a nicely named, easily-testable helper; consider adding a tiny unit test in the existing mod tests block (one case for view, one for non-view) so the policy is locked in without spinning up the integration harness.

[claude]

Claude finished @ovr's task in 37s —— View job

---

Code Review

Verdict: LGTM — focused fix with a tight regression test. 0 high, 0 medium, 2 low.

• Read changed files
• Analyze fallback logic in multi_fact_join_groups.rs
• Review test + snapshot coverage
• Check edge cases (views, missing cubes, errors)

Summary

The fix in multi_fact_join_groups.rs:208-222 correctly addresses the root cause described in the PR body: a dependency-free measure expression (COUNT(*)) resolves to empty hints, and with no dimensions/filters to seed the join the planner used to call JoinGraph.buildJoin([]) and emit a null JoinDefinitionStatic that the bridge couldn't deserialize. Seeding the fallback from the measure's owning cube is the right correction, and gating it by is_view is exactly the subtlety that needs to be there — views are not joinable cubes and their join comes from the other query members.

The new integration test test_expr_measure_count_star_no_hints (member_expressions.rs:147-172) hits the exact scenario from the PR description: a single measure expression with no dimensions/filters and no member dependencies. The snapshot is minimal (SELECT count(*) FROM orders), which is the right output.

Low-severity notes

1. .ok() silently swallows lookup errors (multi_fact_join_groups.rs:246-258)

let is_view = query_tools
    .cube_evaluator()
    .cube_from_path(cube_name.clone())
    .ok()
    .and_then(|cube| cube.static_data().is_view)
    .unwrap_or(false);

If cube_from_path ever fails for a reason other than "cube doesn't exist", this silently defaults to is_view = false and emits a JoinHintItem::Single(<bad name>). The next join_for_hints call will then surface a much less obvious error than the original lookup. Not a blocker since current cube_from_path only fails on missing cubes, but propagating with ? would be more defensive — fallback_hints_for_measure already returns Result<JoinHints, CubeError>.

2. View detection happens per-measure (multi_fact_join_groups.rs:208-222)

Each empty-hint measure triggers an independent cube_from_path call. In practice multi-measure queries with all-empty hints are vanishingly rare so this is fine; if it ever becomes hot it can be memoized on State.

Nits

• The doc comment on the helper (lines 239–241) clearly explains the policy — keep that style.
• A tiny unit test on fallback_hints_for_measure (one view, one non-view case) would lock the policy in without spinning up the integration harness, but the integration coverage is sufficient.
  · fix/tesseract-join-hint-member-expression

[claude]

Claude finished @ovr's task in 32s —— View job

---

Code Review

Verdict: LGTM — focused fix with a tight regression test. 0 high, 0 medium, 2 low.

Review details

• Read changed files
• Analyze fallback logic in multi_fact_join_groups.rs
• Review test + snapshot coverage
• Check edge cases (views, missing cubes, errors)

Summary

The fix in multi_fact_join_groups.rs:208-222 correctly addresses the root cause from the PR body: a dependency-free measure expression (COUNT(*)) resolves to empty hints, and with no dimensions/filters to seed the join the planner used to call JoinGraph.buildJoin([]) and emit a null JoinDefinitionStatic that the bridge couldn't deserialize. Seeding the fallback from the measure's owning cube is the right correction, and gating it by is_view is the subtlety that needs to be there — views are not joinable cubes and their join comes from other query members.

The new integration test test_expr_measure_count_star_no_hints (member_expressions.rs:147-172) targets exactly the scenario from the PR description, and the snapshot (SELECT count(*) FROM orders) is minimal.

Low-severity notes

1. .ok() silently swallows lookup errors (multi_fact_join_groups.rs:246-251)

let is_view = query_tools
    .cube_evaluator()
    .cube_from_path(cube_name.clone())
    .ok()
    .and_then(|cube| cube.static_data().is_view)
    .unwrap_or(false);

If cube_from_path ever fails for a reason other than "cube doesn't exist", this silently defaults to is_view = false and emits a JoinHintItem::Single(<bad name>). The next join_for_hints call then surfaces a much less obvious error than the original lookup. Not a blocker since current cube_from_path only fails on missing cubes, but propagating with ? would be more defensive — fallback_hints_for_measure already returns Result<JoinHints, CubeError>.

2. View detection happens per-measure (multi_fact_join_groups.rs:208-222)

Each empty-hint measure triggers an independent cube_from_path call. In practice multi-measure queries with all-empty hints are vanishingly rare so this is fine; if it ever becomes hot it can be memoized on State.

Nits

• The doc comment on the helper (lines 239–241) clearly explains the policy — keep that style.
• A tiny unit test on fallback_hints_for_measure (one view, one non-view case) would lock the policy in without spinning up the integration harness, but the integration coverage is sufficient.

· `fix/tesseract-join-hint-member-expression`

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 79.12%. Comparing base (bd8b6e6) to head (0aaf4ab).

Additional details and impacted files

@@            Coverage Diff             @@
##           master   #11136      +/-   ##
==========================================
- Coverage   83.63%   79.12%   -4.51%     
==========================================
  Files         256      473     +217     
  Lines       79019    96514   +17495     
  Branches        0     3558    +3558     
==========================================
+ Hits        66085    76368   +10283     
- Misses      12934    19628    +6694     
- Partials        0      518     +518     

Flag	Coverage Δ
cube-backend	58.74% <ø> (?)
cubesql	83.63% <ø> (+<0.01%)	⬆️

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

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.