How to get count of two-way combinations from two columns?

问题

I work for a trucking company, and we're interested in getting a count of the number of times one of our trucks travels between two cities, in either direction.

I have a table that lists an origin and destination for each trip segment, such as:

Origin    Destination

City 1    City 2
City 2    City 1
City 3    City 4
City 2    City 1

I need a query that tells me that there were three trips between City 1 and City 2 and one trip between City 3 and City 4. Thanks very much!

回答1:

I think the following should do the trick.

SELECT route , COUNT(1) FROM 
(
   SELECT 
   CASE WHEN Origin > Destination THEN Origin+'_'+Destination
   ELSE Destination+'_'+Origin 
   END AS route
   FROM table1
 )a
 GROUP BY route

回答2:

Here is a way by sorting the columns, sort of. Granted, I'd change the column names from the output to something other than Origin and Destination since were are basically turning it into a "route". Perhaps just concat them like a1ex07 did. I left it normalized in case you wanted to alter it.

declare @table table (Origin varchar(16), Destination varchar(16))
insert into @table
values
('City 1','City 2'),
('City 2','City 1'),
('City 3','City 4'),
('City 2','City 1')

;with cte as(
select 
    case when Origin > Destination then Origin else Destination end as Origin
    ,case when Destination < Origin then Destination else Origin end as Destination
from
    @table)

select
    Origin
    ,Destination
    ,count(Origin + Destination) 
from 
    cte
group by
    Origin
    ,Destination

回答3:

As they say, if you have right data structures, you usually get the right algorithms for free.

I'm guessing that your schema contains tables similar to the following:

create table City
(
    id int primary key identity,
    name nvarchar(100) not null
)

create table TravelLog
(
    trip_id int primary key identity,
    origin int foreign key references City,
    destination int foreign key references City,
    check (origin <> destination)
)

So, if you add the following two fields to your TravelLog table:

alter table TravelLog
    add 
        low as case when origin <= destination then origin else destination end persisted,
        high as case when origin >= destination then origin else destination end persisted

Then you can use the following simple query to get what you want:

with Q as
(
    select count (*) as trips, low, high from TravelLog group by low, high
)

select Q.trips, o.name point_A, d.name point_B from Q 
    inner join City o on o.id = Q.low
    inner join City d on d.id = Q.high

As a side benefit, you can use that same query to filter by driver, date, etc.

来源：https://stackoverflow.com/questions/45578075/how-to-get-count-of-two-way-combinations-from-two-columns

标签

sql-server

tsql

aggregate-functions