How to bucket a SQL server query

Question

I have the following table in SQL Server:

Date            Browser         Country         Time(ms)
----------------------------------------------------------
201         


        

Answer 1

With conditional aggregation you already were on the right track for one possible solution. But have the condition on the country. In a WHERE filter for the durations and stick the different durations together with UNION ALL.

SELECT '0 - 1s' [Time(ms)],
       count(CASE
               WHEN [Country] = 'US' THEN
                 1
             END) [US],
       count(CASE
               WHEN [Country] = 'JP' THEN
                 1
             END) [JP]
       FROM [dbo].[mytable]
       WHERE [Time] >= 0
             AND [Time] < 1000
UNION ALL
...
UNION ALL
SELECT '+3s' [Time(ms)],
       count(CASE
               WHEN [Country] = 'US' THEN
                 1
             END) [US],
       count(CASE
               WHEN [Country] = 'JP' THEN
                 1
             END) [JP]
       FROM [dbo].[mytable]
       WHERE [Time] >= 3000;

Answer 2

The following will give you a single list which includes the country i.e. so it doesn't make each country a column. To make each a column you either have to do what Sticky-Bit has done, or you have to pivot - either way you have to handle each country individually (unless you build the pivot using dynamic SQL). But maybe your end user can use this list instead of needing a separate column?

declare @MyTable table ([Date] date, Browser varchar(32), Country varchar(2), [Time] int)

insert into @MyTable ([Date], Browser, Country, [Time])
  select '2019-05-06',      'Chrome',          'US',              1000
  union all select '2019-05-06',      'Chrome',          'US',              560
  union all select '2019-05-07',      'Firefox',         'JP',              2300
  union all select '2019-05-07',      'Edge',            'US',              1200
  union all select '2019-05-07',      'Chrome',          'JP',              3000
  union all select '2019-05-07',      'Chrome',          'JP',              3200
  union all select '2019-05-07',      'Chrome',          'JP',              2100
  union all select '2019-05-07',      'Firefox',         'US',              2200

select Duration, Country, count(*)
from (
  select *
    , CASE WHEN [Time] >= 0 AND [Time] < 1000 THEN '0 - 1s'
    WHEN [Time] >= 1000 AND [Time] < 2000 THEN '1 - 2s'
    WHEN [Time] >= 2000 AND [Time] < 3000 THEN '2 - 3s'
    ELSE '+3s' END Duration
    , CASE WHEN [Time] >= 0 AND [Time] < 1000 THEN 0
    WHEN [Time] >= 1000 AND [Time] < 2000 THEN 1
    WHEN [Time] >= 2000 AND [Time] < 3000 THEN 2
    ELSE 3 END DurationOrder
  from @MyTable
) X
group by Duration, DurationOrder, Country
order by DurationOrder, Country