[Bug] Regression in 5.0.0: TypeError "Cannot mix BigInt and other types" when rendering charts with large integers

[jongkookson]

Bug description

Describe the bug
When visualizing data that contains large integer values, the frontend throws a TypeError: Cannot mix BigInt and other types.

It seems that Superset correctly identifies these large numbers as BigInt in the data fetching layer, but the visualization library (ECharts) or the transformation layer attempts to mix these BigInt values with standard Number types during rendering, causing the crash.

This issue was NOT present in Superset version 4.1.4 (and earlier). It appears to be a regression introduced in version 5.0.0.

Reproduction Steps

1. Go to SQL Lab.
2. Select a database (e.g., PostgreSQL).
3. Run the following query to generate virtual data with large integers (approx. 9~10 quadrillion):

   SELECT '2025-01-01'::date date_key, 10402176025875170 a, 9182221549972519 b
   UNION ALL
   SELECT '2025-01-02'::date date_key, 290783677762650 a, 9915708291358753 b

4. Click Create Chart and select Line Chart (or Bar Chart).
5. Set X-axis to date_key.
6. Set Metrics to sum(a) and sum(b).
7. With these settings, the generated query looks like this:

   SELECT
       date_key AS date_key,
       sum(a) AS "sum(a)",
       sum(b) AS "sum(b)"
   FROM (
       SELECT '2025-01-01'::date date_key, 10402176025875170 a, 9182221549972519 b
       UNION ALL
       SELECT '2025-01-02'::date date_key, 290783677762650 a, 9915708291358753 b
   ) AS virtual_table
   GROUP BY date_key
   ORDER BY "sum(a)" DESC
   LIMIT 1000;

8. Click Update Chart.
9. Observe the error.

Expected behavior
The chart should render correctly, similar to how it behaved in version 4.1.4. The frontend should ideally cast BigInt to Number automatically (even with a precision loss warning) or handle large integers gracefully within the visualization library, rather than crashing the entire component.

Screenshots/recordings

Superset version

5.0.0

Python version

3.10

Node version

I don't know

Browser

Chrome

Additional context

This issue was NOT present in Superset version 4.1.4 (and earlier). It appears to be a regression introduced in version 5.0.0.

Checklist

• I have searched Superset docs and Slack and didn't find a solution to my problem.
• I have searched the GitHub issue tracker and didn't find a similar bug report.
• I have checked Superset's logs for errors and if I found a relevant Python stacktrace, I included it here as text in the "additional context" section.

[dosubot]

Hi @jongkookson! I'm Dosu and I’m helping the superset team.

This is a confirmed regression in Superset 5.0.0: mixing BigInt and Number types in the frontend (especially with ECharts) causes a TypeError when rendering charts with very large integers. This was not present in 4.1.4 and earlier.

Recent fixes in Superset (after August 2025) address BigInt serialization and rendering by converting BigInt values to strings before comparison and serialization in filters and charts, and by improving backend JSON utilities to handle BigInt safely (details). These fixes are not included in 5.0.0.

For Superset 5.0.0, the most reliable workaround is to cast large integer fields to string in your SQL queries, for example: SELECT CAST(a AS TEXT) AS a, CAST(b AS TEXT) AS b .... This avoids the frontend mixing BigInt and Number types. Some users report partial success by casting to BIGINT, but if the frontend still receives mixed types, the error can persist (discussion).

If possible, upgrade to a Superset commit after August 2025 to get the latest BigInt handling fixes directly. If you need a frontend patch, you would need to ensure all large values are stringified before rendering or serialization.

There is no documented configuration workaround beyond casting to string or BIGINT in SQL. Persistent issues may require frontend code modification or upgrading to a version with the relevant fixes.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}