DB2，用group by语句数行

Question

I have a pretty simple query in db2 that returns 3 different conditional counts, each as a value:

SELECT
    IFNULL(COUNT(CASE WHEN CURRENT DATE BETWEEN n.start_date AND n.expire_date
                      THEN 1 END), 0) AS current_year,
    IFNULL(COUNT(CASE WHEN CURRENT DATE -  1 YEAR
                          BETWEEN n.start_date AND n.expire_date
                      THEN 1 END), 0) AS prior_year,
    IFNULL(COUNT(CASE WHEN DATE('2018-12-31') - 7 DAY
                          BETWEEN n.start_date AND n.expire_date
                      THEN 1 END), 0) AS full_year
FROM newData n
WHERE n.customer = 111

Which works fine and currently returns something like:

current_year | prior_year  |  full_year
---------------------------------------
412             562             613

The problem is, I want to group them by certain conditions but it no longer returns the one result set, rather a lot of result sets each with their own count:

SELECT
    IFNULL(COUNT(CASE WHEN CURRENT DATE BETWEEN n.start_date AND n.expire_date
                      THEN 1 END), 0) AS current_year,
    IFNULL(COUNT(CASE WHEN CURRENT DATE -  1 YEAR
                          BETWEEN n.start_date AND n.expire_date
                      THEN 1 END), 0) AS prior_year,
    IFNULL(COUNT(CASE WHEN DATE('2018-12-31') - 7 DAY
                          BETWEEN n.start_date AND n.expire_date
                      THEN 1 END), 0) AS full_year
FROM newData n
WHERE n.customer = 111
GROUP BY color,category

So rather than a count of every line I'd like to get a count of the groupings (basically lowering each number/count but still only result set)

Is there something that i'm missing to properly do this with?

UPdate:

An example -

If I have the following rows

product  |  color  |  category  |  start_date  |  expire_date
-------------------------------------------------------------
1            Red        Leather   '2018-12-01'    '2018-12-31'
2            Red        Leather   '2018-12-01'    '2018-12-31'
3            Red        Leather   '2018-12-01'    '2018-12-31'
4            Blue        Leather   '2018-12-01'    '2018-12-31'
5            Blue        Leather   '2018-12-01'    '2018-12-31'
6            Blue        Leather   '2018-12-01'    '2018-12-31'
7            Green        Leather   '2018-12-01'    '2018-12-31'
8            Green        Leather   '2018-12-01'    '2018-12-31'
9            Green        Leather   '2018-12-01'    '2018-12-31'
10           Green        Leather   '2018-12-01'    '2018-12-31'

So if they all fall within a certain date range, say prior_year, then I would want that count to be 3 instead of 10

If I did a select similar to above

select
    ifnull(count(case when '2018'12'15' between start_date and expire_date then 1 end),0) as LastYear
from newData
group by color,category

I would expect:

lastYear
--------
   3

Answer 1

We can try aggregating twice here, once to find the groups which meet your requirements, and the second to take the total counts over all groups.

WITH cte AS (
    SELECT
        detail1,
        detail2,
        detail3,
        COUNT(*) AS cnt,
        COUNT(CASE WHEN CURRENT DATE BETWEEN n.start_date AND n.expire_date
                   THEN 1 END) AS current_year,
        COUNT(CASE WHEN CURRENT DATE -  1 YEAR BETWEEN n.start_date AND n.expire_date
                   THEN 1 END)AS prior_year,
    COUNT(CASE WHEN DATE('2018-12-31') - 7 DAY
        BETWEEN n.start_date AND n.expire_date THEN 1 END) AS full_year
    FROM newData n
    WHERE n.customer = 111
    GROUP BY
        detail1,
        detail2,
        detail3
)

SELECT
    COUNT(CASE WHEN cnt = current_year THEN 1 END) AS current_year,
    COUNT(CASE WHEN cnt = prior_year   THEN 1 END) AS prior_year,
    COUNT(CASE WHEN cnt = full_year    THEN 1 END) AS full_year
FROM cte;