Question

我有以不同价格出售的产品。我想看看在特定价格范围内售出了多少产品。为此，我需要浏览数据，看看如何划分范围，然后获得该范围内的产品计数。

数据如下所示-

Product        Price sold
   A              4.5
   B              45.7
   C              20
   D              20.1
   E              36.8
   F              50

例如，对于以上数据，我看到最小值为4.5，最大值为50。因此，我决定将价格范围划分为-0-10 $，11-20 $，21-30 $，30- 40 $，40-50 $

因此，结果应类似于-

Range           No. of products sold
0-10                    1
11-20                   2
21-30                   0
30-40                   1
40-50                   2

价格为浮动价格，因此，范围应考虑浮动值。这可能吗？

Answer 1

您可以使用generate_array()。我可以这样说：

select lb, lb + 10 as ub, count(d.product)
from unnest(generate_array(0, 50, 10)) lb left join
     data d
     on d.price >= lb and
        d.price < lb + 10
group by lb
order by lb;

您可以将下限和上限连接在一起，但是将它们保留在两列中似乎很有用。

Answer 2

以下是BigQuery标准SQL

#standardSQL
WITH price_ranges AS (
  SELECT '0-10' price_range UNION ALL 
  SELECT '11-20' UNION ALL 
  SELECT '21-30' UNION ALL 
  SELECT '30-40' UNION ALL 
  SELECT '40-50' 
)
SELECT price_range, COUNT(1) number_sold
FROM `project.dataset.table`
JOIN price_ranges 
ON CAST(price_sold AS INT64)
  BETWEEN CAST(SPLIT(price_range, '-')[OFFSET(0)] AS INT64) 
  AND CAST(SPLIT(price_range, '-')[OFFSET(1)] AS INT64)
GROUP BY price_range
-- ORDER BY price_range

如果要应用于问题的样本数据-结果为

Row price_range number_sold  
1   0-10        1    
2   11-20       2    
3   30-40       1    
4   40-50       2

Answer 3

当前答案似乎都无法回答以下问题：“我如何生成范围”（因为两个答案均假设范围为0-50）。

您似乎想要的是一个直方图，您可以在这里找到答案：

https://stackoverflow.com/a/56447260/132438

现在，如果您想在每个存储桶之间进行逐步操作：


WITH data AS (
  SELECT * FROM  `fh-bigquery.public_dump.gdp_capita`
), min_and_max AS (
  SELECT MIN(gdp_capita) min, MAX(gdp_capita) max
  FROM data
), generate_buckets AS (
  SELECT x bucket_min
    , IFNULL(LEAD(x) OVER(ORDER BY x), 1+(SELECT max FROM min_and_max)) bucket_max 
  FROM UNNEST(generate_array(
    (SELECT 0 FROM min_and_max) # min or 0, depending on your start
    , (SELECT max FROM min_and_max)
    , (SELECT POW(10, fhoffa.x.int(LOG10(max-min)))/10 FROM min_and_max) # log10 for round order of 10 steps
  )) x
)

SELECT *
FROM generate_buckets

有了这些水桶，您现在就可以得到直方图：

SELECT bucket_min, bucket_max, COUNT(*) c
FROM generate_buckets
JOIN data
ON data.gdp_capita >= bucket_min AND data.gdp_capita < bucket_max 
GROUP BY 1,2 
ORDER BY 1

如果您还需要具有0个元素的存储桶：

SELECT * REPLACE(IFNULL(c,0) AS c)
FROM (
  SELECT bucket_min, bucket_max, COUNT(*) c
  FROM generate_buckets
  JOIN data
  ON data.gdp_capita >= bucket_min AND data.gdp_capita < bucket_max AND data.one=generate_buckets.one
  GROUP BY 1,2 
  ORDER BY 1
) 
RIGHT JOIN generate_buckets USING(bucket_min, bucket_max)

根据数据创建范围

3 个答案: