count trip in sql server - sql

Question

my table structure
id zoneid status
1 35 IN starting zone
2 35 OUT 1st trip has been started
3 36 IN
4 36 IN
5 36 OUT
6 38 IN last station zone 1 trip completed
7 38 OUT returning back 2nd trip has start
8 38 OUT
9 36 IN
10 36 OUT
11 35 IN when return back in start zone means 2nd trip complete
12 35 IN
13 35 IN
14 35 OUT 3rd trip has been started
15 36 IN
16 36 IN
17 36 OUT
18 38 IN 3rd trip has been completed
19 38 OUT 4th trip has been started
20 38 OUT
21 36 IN
22 36 OUT
23 35 IN 4th trip completed
24 35 IN
now i want a sql query, so i can count no of trips. i do not want to use status field for count
edit
i want result total trips
where 35 is the starting point and 38 is the ending point(this is 1 trip), when again 35 occures after 38 means 2 trip and so on.

Answer 1

So you don't want to look at the status, but only look at the zoneid changes ordered by id. zoneid 36 is irrelevant, so we select 35 and 38 only, order them by id and count changes. We detect changes by comparing a record with the previous one. We can look into a previous record with LAG.
select sum(ischange) as trips_completed
from
(
select
case when zoneid <> lag(zoneid) over (order by id) then 1 else 0 end as ischange
from trips
where zoneid in (35,38)
) changes_detected;

Answer 2

I am suggesting this without any testing. Does the following query produce the correct number of rows? Note if there is a date_created (datetime) column then I would suggest using that column to order by instead of id.
select
ca.in_id, t.id as out_id, ca.in_status, t.status as out_status
from table1 t
cross apply (
select top (1) id as in_id, status as in_status
from table1
where table1.id < t.id
and zoneid = 35
order by id DESC
) ca
where t.zoneid = 38
/* and conditions for selecting one day only */
If that logic is correct then just use COUNT(*) instead of the column list.
CREATE TABLE Table1
("id" int, "zoneid" int, "status" varchar(3), "other" varchar(54))
;
INSERT INTO Table1
("id", "zoneid", "status", "other")
VALUES
(1, 35, 'IN', 'starting zone'),
(2, 35, 'OUT', '1st trip has been started'),
(3, 36, 'IN', NULL),
(4, 36, 'IN', NULL),
(5, 36, 'OUT', NULL),
(6, 38, 'IN', 'last station zone 1 trip completed'),
(7, 38, 'OUT', 'returning back 2nd trip has start'),
(8, 38, 'OUT', NULL),
(9, 36, 'IN', NULL),
(10, 36, 'OUT', NULL),
(11, 35, 'IN', 'when return back in start zone means 2nd trip complete'),
(12, 35, 'IN', NULL),
(13, 35, 'IN', NULL),
(14, 35, 'OUT', '3rd trip has been started'),
(15, 36, 'IN', NULL),
(16, 36, 'IN', NULL),
(17, 36, 'OUT', 'other'),
(18, 38, 'IN', '3rd trip has been completed'),
(19, 38, 'OUT', '4th trip has been started'),
(20, 38, 'OUT', NULL),
(21, 36, 'IN', NULL),
(22, 36, 'OUT', NULL),
(23, 35, 'IN', '4th trip completed'),
(24, 35, 'IN', NULL)
;

Answer 3

For learning purposes, here is a self explanatory & detailed version http://sqlfiddle.com/#!15/d8bf4/1/0
The solution is based on calculating the 'running' count of trips from '35 to 38' and '38 to 35'. Solution is very specific to the OP query but can be optimized with a much shorter version...
with trip_38_to_35 as (
select * from zonecount
where (zoneid=38 and status='OUT') OR (zoneid=35 and status='IN')
order by id asc
)
, count_start_on_38 as (
select count(*) as start_on_38
from trip_38_to_35
where (zoneid=38 and status='OUT') AND
id <
( select max(id)
from trip_38_to_35
where (zoneid=35 and status='IN')
) /*do not count unfinished trips*/
)
, count_end_on_35 as (
select count(*) as end_on_35
from trip_38_to_35
where (zoneid=35 and status='IN')
) /*the other way of trip*/
, trip_35_to_38 as (
select * from zonecount
where (zoneid=35 and status='OUT') OR (zoneid=38 and status='IN')
order by id asc
)
,count_start_on_35 as (
select count(*) as start_on_35
from trip_35_to_38
where (zoneid=35 and status='OUT') AND
id <
( select max(id)
from trip_35_to_38
where (zoneid=38 and status='IN')
) /*do not count unfinished trips*/
)
,count_end_on_38 as (
select count(*) as end_on_38
from trip_35_to_38
where (zoneid=38 and status='IN')
)
/*sum the MIN of the two trips count*/
select
(case when end_on_35 > start_on_38 then start_on_38 else end_on_35 end) +
(case when end_on_38 > start_on_35 then start_on_35 else end_on_38 end)
from
count_start_on_38,
count_end_on_35,
count_start_on_35,
count_end_on_38
btw, 6 trips are calculated as per definition