datafuselabs / wizard

A GPT4 powered tool for detecting bugs in Databend

wizard

A GPT-4 powered tool for detecting bugs in Databend.

wizard's Issues

SELECT-W3 result not stable

SELECT-W3 order is not stable, need to fix.

Run result:

DIFFERENCE FOUND

SELECT-W3:



-- SELECT-W3: Determine the growth in sales quantity for each product from the first sale to the latest sale, with stable ordering
SELECT product_id,
       first_sale_quantity,
       last_sale_quantity,
       last_sale_quantity - first_sale_quantity AS growth
FROM (
         SELECT product_id,
                FIRST_VALUE(quantity) OVER (PARTITION BY product_id ORDER BY sale_date ASC) AS first_sale_quantity,
                 LAST_VALUE(quantity) OVER (PARTITION BY product_id ORDER BY sale_date ASC RANGE BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING) AS last_sale_quantity
         FROM sales
     ) AS sub
ORDER BY growth DESC, product_id ASC
    LIMIT 10
Differences:

bendsql:
1406	1	21	20
1406	1	21	20
1406	1	21	20
1406	1	21	20
1406	1	21	20
1406	1	21	20
1406	1	21	20
1406	1	21	20
1406	1	21	20
1406	1	21	20

snowsql:
2550	1	21	20
2550	1	21	20
2550	1	21	20
2550	1	21	20
2550	1	21	20
2550	1	21	20
2550	1	21	20
2550	1	21	20
2550	1	21	20
2550	1	21	20

SELECT-W12 order not stable

SELECT-W12:


-- SELECT-W12: Display the top 5 products with the highest average sales quantity, along with their rank across all categories
WITH ProductAverage AS (
    SELECT
        p.product_id,
        p.product_name,
        AVG(s.quantity) AS avg_quantity
    FROM
        products p
            JOIN
        sales s ON p.product_id = s.product_id
    GROUP BY
        p.product_id, p.product_name
)
SELECT
    product_id,
    product_name,
    TRUNCATE(avg_quantity, 2),
    RANK() OVER (ORDER BY avg_quantity DESC) AS overall_rank
FROM
    ProductAverage
ORDER BY
    overall_rank
    LIMIT 5
Differences:

bendsql:
29952	Product 29952	14.18	1
18738	Product 18738	14.08	2
32378	Product 32378	14.08	2
26567	Product 26567	14.06	4
30774	Product 30774	14.06	4

snowsql:
29952	Product 29952	14.18	1
32378	Product 32378	14.08	2
18738	Product 18738	14.08	2
26567	Product 26567	14.06	4
30774	Product 30774	14.06	4

SELECT-W13

DIFFERENCE FOUND

SELECT-W13:


-- SELECT-W13: Calculate a cumulative total of sales and a running three-month average, then rank these by customer
WITH SalesData AS (
    SELECT
        customer_id,
        sale_date,
        net_paid,
        SUM(net_paid) OVER (PARTITION BY customer_id ORDER BY sale_date) AS cumulative_sales,
            AVG(net_paid) OVER (PARTITION BY customer_id ORDER BY sale_date ROWS BETWEEN 2 PRECEDING AND CURRENT ROW) AS running_3m_avg
    FROM
        sales
)
SELECT
    customer_id,
    sale_date,
    cumulative_sales,
    running_3m_avg,
    RANK() OVER (ORDER BY running_3m_avg DESC, cumulative_sales DESC) AS sales_rank
FROM
    SalesData
ORDER BY
    customer_id, sale_date
    LIMIT 10
Differences:

bendsql:
0	2021-04-24	14.14	14.1400	800148
0	2021-04-25	22.74	11.3700	1844037
0	2021-05-07	35.11	11.7033	1678601
0	2021-07-25	46.51	10.7900	2117114
0	2021-09-24	54.73	10.6633	2172949
1	2021-01-07	7.36	7.3600	3691368
1	2021-03-23	22.78	11.3900	1833730
1	2021-04-23	42.45	14.1500	797047
1	2021-11-16	51.35	14.6633	678086
1	2021-11-22	65.86	14.3600	745871

snowsql:
0	2021-04-24	14.14	14.14000	800164
0	2021-04-25	22.74	11.37000	1844037
0	2021-05-07	35.11	11.70333	1678589
0	2021-07-25	46.51	10.79000	2117106
0	2021-09-24	54.73	10.66333	2172924
1	2021-01-07	7.36	7.36000	3691305
1	2021-03-23	22.78	11.39000	1833736
1	2021-04-23	42.45	14.15000	797060
1	2021-11-16	51.35	14.66333	678100
1	2021-11-22	65.86	14.36000	745878