SQL查询显示连续3行或更多行且人数超过100人的记录

Question

I have a table with the following structure:

+----+------------+--------+
| id | visit_date | people |
+----+------------+--------+
|  1 | 2017-01-01 |     10 |
|  2 | 2017-01-02 |    109 |
|  3 | 2017-01-03 |    150 |
|  4 | 2017-01-04 |     99 |
|  5 | 2017-01-05 |    145 |
|  6 | 2017-01-06 |   1455 |
|  7 | 2017-01-07 |    199 |
|  8 | 2017-01-08 |    188 |
+----+------------+--------+

I need a query to display the records which have 3 or more consecutive rows and the amount of people more than 100. So the output would be:

+----+------------+--------+
| id | visit_date | people |
+----+------------+--------+
|  5 | 2017-01-05 |    145 |
|  6 | 2017-01-06 |   1455 |
|  7 | 2017-01-07 |    199 |
|  8 | 2017-01-08 |     88 |
+----+------------+--------+

Create table statement for convenience:

CREATE TABLE stadium(id INT, visit_date DATE, people INT);
INSERT INTO stadium VALUES(1, '01/JAN/17', 10);
INSERT INTO stadium VALUES(1, '02/JAN/17', 109);
INSERT INTO stadium VALUES(1, '03/JAN/17', 150);
INSERT INTO stadium VALUES(1, '04/JAN/17', 99);
INSERT INTO stadium VALUES(1, '05/JAN/17', 145);
INSERT INTO stadium VALUES(1, '06/JAN/17', 1455);
INSERT INTO stadium VALUES(1, '07/JAN/17', 199);
INSERT INTO stadium VALUES(1, '08/JAN/17', 188);

This is in Oracle SQL

I would highly obliged if anyone can explain the logic along with solution.

Answer 1

This is a type of gaps-and-islands problem. You can use the difference of row numbers to identify the groups. Then a count(*) to count the rows and filter:

select s.*
from (select s.*,
             count(*) over (partition by id, seqnum - seqnum_2) as num_in_row
      from (select s.*,
                   row_number() over (partition by id order by visit_date) as seqnum,
                   row_number() over (partition by id, (case when people > 100 then 1 else 0 end) order by visit_date) as seqnum_2
            from stadium s
           ) s
      where people > 100
     ) s
where num_in_row >= 3;

Here is SQL Fiddle.

Answer 2

This is a typical pattern matching problem. match_recognize does quick work of such tasks.

select id, visit_date, people
from   stadium
match_recognize(
  order by visit_date
  all rows per match
  pattern ( a{3, } )
  define a as people >= 100    --  or > 100?
);