我需要使用 google bigquery 中 iowa_liquor_sales_forecasting 数据集上的 2 个表知道 2020 年到 2021 年的销售额增长率

Question

我正在使用此查询，但没有要显示的数据。我的查询有什么问题？

with t1 as
(
    select
        extract (year from date) as year,
        sum (sale_dollars) as sales_year0
    from `bigquery-public-data.iowa_liquor_sales_forecasting.2020_sales_train`
    group by 1
),
t2 as
(
    select
        extract (year from date) as year,
        sum (sale_dollars) as sales_year1
    from `bigquery-public-data.iowa_liquor_sales_forecasting.2021_sales_predict`
    group by 1
)
select *,
    round((sales_year1 - lag(sales_year0) over(order by t1.year asc)) / lag(sales_year0) over(order by t1.year asc) * 100,2)||'%' as growth_rate
from t1
join t2 on t1.year=t2.year

我的预期回报

year | total_sales | growth_rate |
2020 | ........... | ........... |
2021 | ........... | ........... |

Answer 1

BigQuery 中的JOIN 本身（没有left，outer，...）是inner join，这意味着左边的表和右边的表必须共享一个条件才能产生结果。

目前，t1 仅具有 2020 年，t2 仅具有 2021 年，这导致连接条件 join t2 on t1.year=t2.year 为假。

FROM A [INNER] 加入 B 使用 (x)

我猜你想要做的是计算growth_rate，下面是获取价值的一种方法。

with
t1 as (
    select
        extract (year from date) as year_2020,
        sum (sale_dollars) as sales_year_2020
    from `bigquery-public-data.iowa_liquor_sales_forecasting.2020_sales_train`
    group by 1
),
t2 as (
    select
        extract (year from date) as year_2021,
        sum (sale_dollars) as sales_year_2021
    from `bigquery-public-data.iowa_liquor_sales_forecasting.2021_sales_predict`
    group by 1
)
select
    *,
    -- round((year_2021 - year_2020) / year_2020 * 100, 3) as growth_rate  -- typo (wrong column names)
    round((sales_year_2021 - sales_year_2020) / sales_year_2020 * 100, 3) as growth_rate
from t1 join t2 on t1.year_2020 = (t2.year_2021 - 1)
;