PR Summary

Low Risk
Low risk, localized query change that only affects how rowCount is computed when fetching a table; main concern is potential query/perf differences due to the new join/grouping.

Overview
Fixes incorrect table rowCount when fetching a single table by switching getTableById to compute the count directly in SQL.

getTableById now selects explicit table fields, LEFT JOINs user_table_rows, and uses count() + groupBy to return an accurate rowCount (including 0 for empty tables), instead of relying on a stale/unmaintained value.

^{Written by Cursor Bugbot for commit 91fad78. This will update automatically on new commits. Configure here.}

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated (UTC)
docs	Skipped		Mar 17, 2026 5:23am

Greptile Summary

This PR fixes a bug in getTableById where rowCount was sourced from a stored property on the table definition that was never updated, causing it to always be stale. The fix switches to computing rowCount dynamically via a SQL COUNT aggregate over userTableRows, using a LEFT JOIN and GROUP BY, which is identical to the pattern already used in the listTables function.

Key changes:

• getTableById now uses an explicit field projection (instead of SELECT *) to selectively select columns from userTableDefinitions alongside the computed rowCount
• A LEFT JOIN on userTableRows is added so that tables with zero rows still appear in the result
• GROUP BY userTableDefinitions.id is added to support the COUNT aggregate
• rowCount is now COALESCE(COUNT(userTableRows.id), 0), matching the existing pattern in listTables

Confidence Score: 5/5

• This PR is safe to merge — it's a targeted, correct bug fix with no regressions introduced.
• The change is minimal and well-scoped: it replaces a stale stored property with a live SQL aggregate, using the exact same pattern already validated in listTables. The LEFT JOIN correctly handles the zero-row case, GROUP BY is required for the aggregate, and LIMIT 1 is preserved. No types or interfaces were changed, and the return shape of TableDefinition is fully satisfied.
• No files require special attention.

Important Files Changed

Sequence Diagram

sequenceDiagram
    participant Caller
    participant getTableById
    participant DB as Database

    Caller->>getTableById: getTableById(tableId, options)
    getTableById->>DB: SELECT id, name, ..., COALESCE(COUNT(rows.id), 0) AS rowCount<br/>FROM userTableDefinitions<br/>LEFT JOIN userTableRows ON rows.tableId = defs.id<br/>WHERE defs.id = tableId [AND archivedAt IS NULL]<br/>GROUP BY defs.id<br/>LIMIT 1
    DB-->>getTableById: result row with computed rowCount
    getTableById-->>Caller: TableDefinition (with accurate rowCount)

_{Last reviewed commit: 91fad78}