SQL. Подсчет изменений в столбце на группу

Question

Установите код:

CREATE TABLE [Table3] ([Team] nvarchar(10), [Int] int, [Player] nvarchar(10))

INSERT INTO [Table3] VALUES ('Yankees', 3, 'Carl')
INSERT INTO [Table3] VALUES ('Cubs',    2, 'Kevin')
INSERT INTO [Table3] VALUES ('Cubs',    3, 'Steve')
INSERT INTO [Table3] VALUES ('Yankees', 1, 'Tony')
INSERT INTO [Table3] VALUES ('Yankees', 2, 'Tony')
INSERT INTO [Table3] VALUES ('Red Sox', 4, 'Phil')
INSERT INTO [Table3] VALUES ('Red Sox', 5, 'Liam')
INSERT INTO [Table3] VALUES ('Dodgers', 5, 'Brad')
INSERT INTO [Table3] VALUES ('Red Sox', 7, 'Phil')
INSERT INTO [Table3] VALUES ('Red Sox', 6, 'Liam')
INSERT INTO [Table3] VALUES ('Cubs',    5, 'Kevin')
INSERT INTO [Table3] VALUES ('Cubs',    4, 'Steve')
INSERT INTO [Table3] VALUES ('Cubs',    6, 'Steve')
INSERT INTO [Table3] VALUES ('Dodgers', 3, 'Tony')
INSERT INTO [Table3] VALUES ('Dodgers', 4, 'Tony')

Я бы хотел отсортировать по Team, а затем Int. И затем я хотел бы видеть количество раз Player изменений для каждого Team.

Окончательный результат:

Team    | Changes
-------------------
Cubs    | 3
Dodgers | 1
Red Sox | 2
Yankees | 1

Answer 1

Вы просто хотите lag() с агрегацией?

select team, count(*)
from (select t.*,
             lag(player) over (partition by team order by [int]) as prev_player
      from table3 t
     ) t
where prev_player is null or prev_player <> player
group by team;