feat(explain): EXPLAIN result views (Explain/Indexes/Projections/Pipeline/Estimate) + Explain button#37
Merged
Merged
Conversation
…line/Estimate) + Explain button Replace the one-dimensional raw EXPLAIN dump with five views in the data pane: - Explain (default) runs the user's EXPLAIN verbatim as clean TabSeparatedRaw, so arbitrary/complex parameters are honored. - Indexes / Projections derive `indexes=1` / `projections=1` from the inner query. - Pipeline derives `EXPLAIN PIPELINE graph=1` and renders the Graphviz DOT as a zero-dep SVG layered graph (pure parse+layout in src/core/dot.js). - Estimate derives `EXPLAIN ESTIMATE` and renders structured rows as a table. On Run we auto-select a rich view only when the typed statement is exactly that canonical form; anything else stays on the verbatim Explain tab. Switching a view re-runs the derived query and never edits the editor SQL. New "Explain" toolbar button (between Format and Save) explains any query without editing it. No new runtime dependency. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01QGBS74oUsXarGkCRQKEFLu
…essage Reword the empty-Estimate placeholder (a trivial count() is answered from metadata, so ESTIMATE returns no part rows even on a MergeTree table). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01QGBS74oUsXarGkCRQKEFLu
…e graph test - Pipeline graph now lays out top→bottom: sequential stages stack vertically, parallel processors in a stage sit side-by-side horizontally. Edges flow bottom→top-centre; self-loops are filtered. - EXPLAIN views drop the ms/rows/bytes header stats (not meaningful for a plan) to give the five tabs room. - Add a Playwright e2e test that renders a real `EXPLAIN PIPELINE graph = 1` capture from the antalya ontime dataset (fact/dim join + aggregated subquery join, 37 processors / 40 edges) and asserts the vertical-with-parallel-lanes SVG. Fixture: tests/e2e/fixtures/ontime-pipeline-graph.dot. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01QGBS74oUsXarGkCRQKEFLu
Add an Expand button (results header, Pipeline view) that opens the graph in a fullscreen overlay with wheel-zoom around the cursor, drag-pan, +/- and Fit-to-screen controls, and Esc/✕/backdrop close. Makes complex pipelines (which the current layout draws wide) navigable; better automatic layout is deferred to a follow-up PR. - src/core/panzoom.js: pure viewBox algebra (fit/zoom/pan), 100% covered. - src/ui/explain-graph.js: extract buildPipelineSvg; openPipelineFullscreen. - src/ui/results.js: Expand button; icons.js: expand + minus glyphs. - e2e: pipeline.html exposes __openFullscreen; spec drives real wheel/drag/Esc. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01QGBS74oUsXarGkCRQKEFLu
- run(): a typed EXPLAIN now honors exactly what was typed — a plain EXPLAIN always opens the verbatim Explain view instead of inheriting a stale rich view from a previous run/tab. Drop the unused app.state.explainView. (review #1) - dot.parseDot: blank quoted label strings before edge scanning and only keep edges between declared processors, so a `->` inside a label (e.g. a lambda `x -> x + 1`) no longer fabricates phantom nodes/edges. (review #2) - panzoom.zoomBox: guard a degenerate zero-size viewBox (defensive). - overlay: Esc stops propagation so it doesn't also reach the app key handler. - remove now-dead isExplain from core/format.js (superseded by parseExplain). - document the ontime DOT fixture's capture provenance. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01QGBS74oUsXarGkCRQKEFLu
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Replaces the one-dimensional raw EXPLAIN dump with five views in the data pane, plus an Explain toolbar button (between Format and Save) that explains any query without editing it.
TabSeparatedRaw)EXPLAIN indexes = 1 <inner>EXPLAIN projections = 1 <inner>EXPLAIN PIPELINE graph = 1 <inner>EXPLAIN ESTIMATE <inner>EXPLAIN indexes=1, actions=1 …) run exactly as typed and stay on the Explain tab.EXPLAIN ESTIMATE …→ Estimate); else verbatim Explain.… FORMAT <name>on an EXPLAIN bypasses the views (single raw tab).How
src/core/explain.js(parse/detect/build) andsrc/core/dot.js(DOT parse + layered-DAG layout), both 100%-covered.src/ui/explain-graph.js(built with thes()hyperscript). No new runtime dependency — honors CLAUDE.md hard rule 4.run()gains EXPLAIN-mode handling +setExplainView/explainQueryactions;results.jsrenders the five-tab strip and dispatches text/table/graph.Tests & verification
EXPLAIN ESTIMATE …auto-opens Estimate; editor SQL unchanged across switches. (A plaincount()yields no ESTIMATE rows — answered from metadata — and shows a neutral placeholder.)🤖 Generated with Claude Code