test(parity): first parity harness — port MobilityDB 001_set.test.sql#6
Merged
Conversation
Adds test/sql/parity/001_set.test — a 1:1 port of the upstream MobilityDB regression file mobilitydb/test/temporal/queries/001_set.test.sql into DuckDB sqllogic format. Goal is executable, machine-checkable parity between MobilityDB (PostgreSQL extension) and MobilityDuck (DuckDB extension) for the traditional temporal-set types — so the same query runs on both the historical store and the pipeline side. Every query from the source file is mirrored in order. Queries that run on MobilityDuck today are active assertions. Queries that hit an unbound MEOS symbol, a naming divergence, or an output-format divergence are wrapped in `mode skip` with a `# gap:` annotation that names the specific missing binding. Each skip is a tracked backlog item for a follow-up binding PR. Tracked gaps (initial triage, based on src-tree audit — CI may reveal more divergences on the active assertions): - set(ARRAY[GEOMETRY ...]) — geomset list-constructor not registered in src/geo/geoset.cpp (MEOS: geomset_make). - spans(<set_type>) — only spans(<spanset_type>) is registered in src/temporal/spanset.cpp:330; the set-level form is missing. - splitNspans / splitEachNspans — naming divergence. MobilityDuck registers splitNSpans / splitEachNSpans (camelCase, capital S) in src/temporal/span.cpp:313-340; the MobilityDB SQL form uses lowercase "spans". - set_hash / set_hash_extended — not registered in src/temporal/set.cpp (MEOS: set_hash, set_hash_extended). Output-format note: MobilityDB emits PG-locale output (MM-DD-YYYY dates, "Sat Jan 01 00:00:00 2000 PST" timestamps); MobilityDuck emits DuckDB-shaped output (ISO dates, UTC timestamps under TZ=UTC, as the Makefile sets). Expected blocks in this file reflect MobilityDuck's output — parity is in the query surface, not the display layer. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
CI parser requires every `statement error` directive to carry a `----` separator followed by an expected error-message substring (even empty is rejected). Port was missing both on all 9 error assertions; adds MEOS error substrings lifted from the upstream MobilityDB expected-output file so they match what MEOS raises through the MobilityDuck wrapper: - Could not parse (tstzset parse errors) - cannot be negative (asText negative-digits) - cannot be empty (set of empty array) - mixed 2D/3D, non-empty, mixed SRID (geomset constructor errors — in skip block) - strictly positive (splitNspans / splitEachNspans negative count — in skip block) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Previous CI run died after the first `statement error` query — MEOS's default error handler terminates the process on MEOS_ERROR instead of surfacing as a DuckDB exception, so the parse-error test killed the whole test runner before any assertion could be checked. Root cause (MobilityDuck bug, not a parity-file bug): `src/mobilityduck_extension.cpp:143` calls `meos_initialize()` but not `meos_initialize_error_handler(...)`. Follow-up PR: install a MEOS error handler in LoadInternal() that maps MEOS_ERROR to a thrown DuckDB::InvalidInputException — at which point every `statement error` block in this file becomes a real parity assertion in a single edit. For now, wrap all 4 previously-active `statement error` blocks (tstzset parse errors, asText negative digits, empty-array `set(...)` ) in `mode skip` with cross-references to a new top-of-file error-handler note explaining the gap and the one-stroke fix once it lands. The 5 error assertions already inside geomset / splitN skip blocks get the same cross-reference. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
3 tasks
nhungoc1508
pushed a commit
that referenced
this pull request
Apr 25, 2026
MEOS's default error handler writes the message to stderr and calls exit(EXIT_FAILURE) on errlevel == ERROR (PG elog level 21 carried through via postgres.h). That tears down the whole DuckDB process on any invalid user input — a parse error on a malformed tstzset literal, a negative precision on asText, etc. — which (a) is user-hostile in production and (b) makes every `statement error` test kill the test runner before Catch2 can check it. PR #6 surfaced this by trying to exercise MEOS error paths in the parity harness and crashing CI on the first parse-error query. This PR installs `MobilityduckMeosErrorHandler` via `meos_initialize_error_handler(...)` in LoadInternal, right after `meos_initialize()`. The handler throws a DuckDB::InvalidInputException on errlevel >= 21 (ERROR/FATAL/PANIC), so MEOS errors surface as ordinary query failures with the MEOS message attached. Lower levels (WARNING/NOTICE/INFO) are silently ignored for now — can be wired to DuckDB's logger in a follow-up if useful. Parity file (`test/sql/parity/001_set.test`) consequences: - Deletes the top-of-file error-handler gap note (now obsolete). - Unwraps the 3 MEOS-error-specific `mode skip` blocks covering tstzset parse errors and asText negative-digits — they run as real `statement error` assertions. - Re-wraps ONE of those (`set('{}'::timestamptz[])`) in a skip block with a new gap note: that query fails at the DuckDB cast layer with a conversion error before MEOS sees it, because `'{}'` is PG array-literal syntax that DuckDB rejects as a string literal. The equivalent DuckDB form `set(ARRAY[]::TIMESTAMPTZ[])` reaches MEOS but the constructor wrapper in src/temporal/set_functions.cpp returns null instead of calling meos_error(...) on empty input — a separate gap, documented for a later follow-up. Verified locally (`TZ=UTC ./build/release/test/unittest "/path/to/test/*"`): 747 assertions pass across 13 test cases (up from 739 pre-handler), no regressions. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This was referenced Apr 25, 2026
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
2 participants
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.
Summary
First parity harness between MobilityDB (PostgreSQL extension, historical store) and MobilityDuck (DuckDB extension, pipeline side). Adds
test/sql/parity/001_set.test— a 1:1 port of the upstream regression filemobilitydb/test/temporal/queries/001_set.test.sqlinto DuckDB sqllogic format.Goal: the long-term objective is that the same SQL query runs on both projects and produces comparable results. This file is the first executable, machine-checkable artifact that makes the gap measurable. Every future binding PR can annotate the skip block it unblocks.
Methodology
TZ=UTC) — parity is in the query surface, not the display layer.mode skipblock per query that hits an unbound MEOS symbol, a naming divergence, or an output-format divergence, with a# gap:annotation naming the specific missing binding.Initial gap triage
Four tracked gaps (based on src-tree audit — CI may reveal additional divergences on the currently-active assertions):
set(ARRAY[GEOMETRY ...])not registered insrc/geo/geoset.cppScalarFunction("set", {LIST(GEOMETRY)}, geomset(), ...)(MEOS:geomset_make)spans(<set_type>)missing — onlyspans(<spanset_type>)exists (src/temporal/spanset.cpp:330)spans(MEOS:set_spans)splitNspans/splitEachNspans(lowercase s); MobilityDuck registerssplitNSpans/splitEachNSpansinsrc/temporal/span.cpp:313-340set_hash/set_hash_extendednot registered insrc/temporal/set.cppset_hash,set_hash_extended)Test plan
test/sql/parity/001_set.testand reports pass/fail per active assertion.003_span.test.sqlusing the same shape.Notes
bigintsetqueries for the set of MobilityDB queries that PR adds will be easy follow-ups under the same parity structure.🤖 Generated with Claude Code