如何优化我的sql查询，需要2分25秒才能运行

Question

I have a table with 3 GB data (It will keep on increasing) and I need to display total sales, top category and top product (Maximum occurrence in the column). Following is the query that's giving me the above mentioned result:

select t.category, 
       sum(t.sale) sales,
        (select product 
        from demo 
        where  category = t.category
        group by product
        order by count(*) desc
        limit 1) top_product
from demo t
group by t.category

The above query takes approximately 2 mins and 25 seconds. I couldn't find any way to optimize it. Is there any other way that someone could recommend?

Example table:

category  product    sale 
C1         P1        10
C2         P2        12
C3         P1        14
C1         P2        15
C1         P1        02
C2         P2        10
C2         P3        22
C3         P1        01
C3         P2        27
C3         P3        02

Output:

category  Top product   Total sales 
    C1         P1        27
    C2         P2        44
    C3         P1        44

Answer 1

Your query could be written like this:

SELECT g1.category, g1.sum_sale, g2.product
FROM (
    SELECT category, SUM(sale) AS sum_sale
    FROM demo
    GROUP BY category
) AS g1
INNER JOIN (
    SELECT category, product, COUNT(*) AS product_count
    FROM demo
    GROUP BY category, product
) AS g2 ON g1.category = g2.category
INNER JOIN (
    SELECT category, MAX(product_count) AS product_count_max
    FROM (
        SELECT category, product, COUNT(*) AS product_count
        FROM demo
        GROUP BY category, product
    ) AS x
    GROUP BY category
) AS g3 ON g2.category = g3.category AND g2.product_count = g3.product_count_max

Basically it tries to find the maximum count(*) per category and from that it calculates the product. It could benefit from appropriate indexes.

Answer 2

A MySQL only hack solution is using GROUP_CONCAT in combination with nested SUBSTRING_INDEX functions to get the first element in an Ordered comma separated string.

It is not an ideal approach ; but it will reduce the number of subqueries required, and may be efficient for your peculiar case.

You will also need to use SET SESSION group_concat_max_len = @@max_allowed_packet; .

We basically determine sales and count of occurrence, for a product and category combination. This result-set is then used as a Derived Table , and we use the Group_concat() hack to determine the product with maximum count in a category.

SET SESSION group_concat_max_len = @@max_allowed_packet;

SELECT 
  dt.category, 
  SUM(dt.sale_per_category_product) AS total_sales, 
  SUBSTRING_INDEX(
    SUBSTRING_INDEX(
      GROUP_CONCAT(dt.product ORDER BY dt.product_count_per_category DESC)
                    , ','
                    , 1
                   )
                 , ','
                 , -1
                ) AS top_product 
FROM 
(
  SELECT 
    category, 
    product, 
    SUM(sale) AS sale_per_category_product, 
    COUNT(*) AS product_count_per_category 
  FROM demo 
  GROUP BY category, product 
) AS dt 
GROUP BY dt.category 

Schema (MySQL v5.7)

| category | total_sales | top_product |
| -------- | ----------- | ------------|
| C1       | 27          | P1          |
| C2       | 44          | P2          |
| C3       | 44          | P1          |

---

View on DB Fiddle