tursodatabase/turso · error

ranked row count mismatch

Error message

ranked row count mismatch

What it means

validate_result compares the rows returned by the benchmark query against a count recomputed from the seeded data predicate. For QueryCase::Ranked, expected rows are capped by the LIMIT 10 of the ranked query and multiplied by the number of query executions; ensure! at perf/fts/src/main.rs:333 fails when the observed row count differs, meaning the engine returned a different result set than the model predicts.

Solutions

1. Re-run with a single query (--queries 1) to isolate whether the mismatch is per-query or aggregate
2. Diff the actual returned rows against ids.filter(predicate).take(10) to see what is missing/extra
3. Check for concurrent writers or MVCC snapshot differences between seeding and validation
4. Update validate_result if the query's LIMIT or predicate legitimately changed

Example fix

// before (assumes no cap)
assert_eq!(rows, expected_rows * queries);
// after (ranked queries are LIMIT 10)
assert_eq!(rows, expected_rows.min(10) * queries);

// compute expectation locally before validating
let expected: usize = (0..n).filter(|&id| id % 2 == 0 || id % 3 == 0).count().min(10) * queries;

match validate_result(...) {
    Err(e) if e.to_string().contains("ranked row count mismatch") => {
        // dump actual vs expected rows for diagnosis
        eprintln!("ranked mismatch: got {rows}, want {expected}");
    }
    other => other?,
}

When it happens

Trigger: Calling validate_result with case == Ranked where rows != expected_rows.min(10) * queries — i.e. the FTS engine returned more or fewer matching rows than the capped expectation, typically indicating a query engine, ranking/LIMIT, or MVCC visibility discrepancy.

Common situations: Running the validation under WAL/MVCC where visibility differs; a Turso FTS bug returning duplicate or missing rows; changing the LIMIT or query text without updating validate_result; concurrency races writing rows between seed and validation.

Related errors

• row count mismatch for
• connections must be positive
• documents must be positive
• ID sum mismatch for
• index contains segment bytes, below requested minimum …

AI-assisted analysis of tursodatabase/turso@8d4a589f8d (2026-09-20). Data as JSON: /api/errors/3eec1581705b010d. Report an issue: GitHub.

Appendix: source

Thrown at perf/fts/src/main.rs:333

}

fn validate_result(
    case: QueryCase,
    documents: usize,
    queries: usize,
    rows: usize,
    sum: i64,
) -> Result<()> {
    let ids = (0..documents).filter(|id| match case {
        QueryCase::Rare => id % 100 == 0,
        QueryCase::Common => true,
        QueryCase::And => id % 6 == 0,
        QueryCase::Or | QueryCase::Ranked => id % 2 == 0 || id % 3 == 0,
        QueryCase::Phrase => id % 200 == 0,
    });
    let expected_rows = ids.clone().count();
    if matches!(case, QueryCase::Ranked) {
        ensure!(
            rows == expected_rows.min(10) * queries,
            "ranked row count mismatch"
        );
    } else {
        ensure!(
            rows == expected_rows * queries,
            "row count mismatch for {case:?}"
        );
        ensure!(
            sum == ids.map(|id| id as i64).sum::<i64>() * queries as i64,
            "ID sum mismatch for {case:?}"
        );
    }
    Ok(())
}

#[cfg(test)]
mod tests {

View on GitHub (pinned to 8d4a589f8d)