feat(parser): wire mysql + sqlite parsers for INNER JOIN queries - #48
Merged
Conversation
![devin-ai-integration[bot]](https://avatars.githubusercontent.com/in/811515?s=60&v=4){kind=small}
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
…solver When a query contains the JOIN keyword, route each SELECT column through the multi-table resolver from parser::joins (landed in the prior PR). Qualified columns now resolve across the alias map and land in QueryDef.returns with source_table populated. Queries without JOIN keep the existing single-table path unchanged — ensure_supported_select_expr still rejects qualified selects there. Lifting that restriction for single-table queries is a separate effort (PR #32 was in flight before this series). Changes: - resolve_return_columns gains a &[TableDef] schema_tables param - JOIN_DETECT_RE gates multi-table resolution - resolve_return_columns_multi_table rejects SELECT * across joins with a clear v1.2 pointer - Parser propagates schema_tables from parse_queries to resolve_return_columns Tests: - 5 new postgres parser unit tests (qualified columns, AS aliases, SELECT * rejection, LEFT JOIN rejection, single-table rejection still active) - Updated the cli test that was asserting the old rejection — it now asserts the JOIN query succeeds (renamed from cli_generate_rejects_ qualified_selects to cli_generate_accepts_multi_table_inner_join) 158 lib + 11 CLI + 2 e2e + 4 typecheck tests all pass. Clippy clean.
…ives Devin flagged that `JOIN_DETECT_RE.is_match(sql)` matches `\bJOIN\b` anywhere in the SQL, which false-triggers on queries where the JOIN lives inside a subquery (e.g. `WHERE id IN (SELECT ... JOIN ...)`). The outer query's single-table SELECT was then misrouted to the multi-table resolver and failed with a confusing error. Fix: route through `parser::joins::has_outer_join`, which scopes the check to the outer FROM body via FROM_CLAUSE_RE. JOINs inside subqueries stay behind the WHERE boundary and no longer trip the outer detection. Also drop the local JOIN_DETECT_RE static (was unused after the switch) and the stale imports. New regression test covers the subquery case end-to-end through parse_queries.
JagritGumber
force-pushed
the
feat/postgres-parser-inner-join
branch
from
April 20, 2026 00:17
06917c9 to
4ab2d7d
Compare
Mirror PR #47 (postgres) for the remaining two dialects. Both mysql.rs and sqlite.rs gain: - A `&[TableDef]` schema_tables param on resolve_return_columns - JOIN_DETECT_RE check that routes to the multi-table resolver path - 3 new unit tests per dialect (qualified columns, SELECT * rejection, LEFT JOIN rejection) Also lift the resolve_return_columns_multi_table helper that lived in postgres.rs into parser/joins.rs as resolve_multi_table_columns, shared across all three dialects to avoid copy-paste. Same goes for the JOIN_DETECT_RE constant — now `pub` from joins.rs. 164 lib + 11 CLI + 2 e2e + 4 typecheck tests all pass. Clippy clean. After this PR all three dialects accept INNER JOIN queries with qualified columns. Codegen consuming source_table is the final step (C16).
JagritGumber
force-pushed
the
feat/mysql-sqlite-parser-inner-join
branch
from
April 20, 2026 00:24
bbb94ba to
c48bc07
Compare
JagritGumber
changed the base branch from
feat/postgres-parser-inner-join
to
main
April 20, 2026 00:43
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.
Summary
Third PR in the JOIN series (C15c+d combined). Mirrors #47 for the remaining two dialects: mysql and sqlite.
Stacked on #47 (`feat/postgres-parser-inner-join`). When #47 merges, GitHub auto-retargets to main.
Why combined into one PR
C15c and C15d in the plan were separate PRs per dialect, but the changes are 90% identical (~50 LOC each, same pattern). Splitting would create two PRs that mostly duplicate each other for review effort. Bundling stays small (~100 LOC of dialect changes) and the symmetry is easy to review side-by-side.
Changes
Shared helper lifted into joins.rs:
mysql.rs + sqlite.rs each:
Test plan
Next
C16 — codegen emits disambiguated row fields when joined columns share names (e.g. both `users.id` and `orgs.id` selected → `users_id` and `orgs_id` in the generated row type).