针对多个类别中的项目的总价格分组

Question

Consider the following data:

Items | Price | Categories
--------------------------
Item1 |    10 | Cat1, Cat2
Item2 |    20 | Cat1, Cat3
Item3 |    15 | Cat1, Cat2
--------------------------
Total |    45

If I group the data according to Categories, following stats come up

Categories | Price
------------------
Cat1       |    45
Cat2       |    25
Cat3       |    20
------------------
Total      |    90

Now if I have to sum up the prices, Actual sum will be 45 but the sum according to categories shown will be different ie 90 . So both representations say different stats but they are correct in their own way.

Looking for answer to this question: How would I represent such a stat. For example the total on Top says 45 but the sum across the categories says 90. I mean, isn't it confusing for the end-user?

Any suggestions of how to tackle this problem other than choosing one category per item?

Answer 1

Below is for BigQuery Standard SQL

#standardSQL
SELECT category, SUM(price) AS price FROM (
  SELECT category, price FROM `project.dataset.table`, 
  UNNEST(SPLIT(categories, ', ')) category UNION ALL
  SELECT 'Total', price FROM `project.dataset.table`
)
GROUP BY category

If to apply to sample data from your question as in below example

#standardSQL
WITH `project.dataset.table` AS (
  SELECT 'Item1' AS item, 10 AS price, 'Cat1, Cat2'  AS categories UNION ALL
  SELECT 'Item2', 20, 'Cat1, Cat3' UNION ALL
  SELECT 'Item3', 15, 'Cat1, Cat2' 
)
SELECT category, SUM(price) AS price FROM (
  SELECT category, price FROM `project.dataset.table`, 
  UNNEST(SPLIT(categories, ', ')) category UNION ALL
  SELECT 'Total', price FROM `project.dataset.table`
)
GROUP BY category   

result is

Row category    price    
1   Cat1        45   
2   Cat2        25   
3   Cat3        20   
4   Total       45     

with correct Total value

Answer 2

You can split the data and then re-aggregate:

with t as (
      select 'Item1' as item, 10 as price, 'Cat1, Cat2'  as categories union all
      select 'Item2', 20, 'Cat1, Cat3' union all
      select 'Item3', 15, 'Cat1, Cat2' union all
      select 'Total', 45, NULL
     )
select coalesce(category, 'Total'), sum(price)
from t cross join
     unnest(split(t.categories, ', ')) category
group by rollup(category)
order by category nulls last;

If you don't need the total, then remove the rolllup .