Feature Availability Matrix — Solution & ExplanationProblems / Feature Availability Matrix / EditorialApproach

Group by plan. One output row per plan.

MAX(CASE ...) per feature column. The CASE maps this feature's rows to 1 and everything else to 0; MAX over the group is then a boolean OR — any hit makes the cell 1. Absent features produce an all-zeros column value naturally, and unrelated features (exports) never match a CASE, so they're harmlessly ignored without a WHERE.

Order by plan.

Key InsightMAX(CASE WHEN ... THEN 1 ELSE 0 END) is the presence pivot — the boolean sibling of the SUM(CASE) amount pivot. The mental model: CASE paints each row, the aggregate collapses the paint per group, and *which* aggregate you pick defines the semantics — SUM counts, MAX detects. Entity-attribute tables (features, permissions, tags) almost always want the MAX form.Hints

Each cell answers 'does ANY row in this plan's group match the feature?' — that's MAX over a 0/1 CASE: one matching row pushes the max to 1.

SUM(CASE ...) would also detect presence but could exceed 1 if the log ever had duplicate rows; MAX is the idiomatic boolean OR.

Features outside the three named columns simply never match any CASE — they need no filtering, they just contribute 0s.

Solve Feature Availability Matrix yourself →