SQL：基于多个记录的查询

Question

The following table has the ID number of people along with cities they have worked in:

PERSON_NO | CITY_NAME
---------------------
1         |    City A
2         |    City B
3         |    City A
3         |    City B
3         |    City C
4         |    City A
4         |    City B
4         |    City C

How would I be able to get the PERSON_NO of all the people who have lived in all three cities, A,B, and C?

I want to return

PERSON_NO
---------
3
4

Thanks, again. I haven't had that much experience with SQL and so I'm not sure what to do.

Answer 1

If you're in SQL Server 2000 or higher you can use INTERSECT:

select PERSON_NO from table1 where CITY_NAME='City A'
intersect
select PERSON_NO from table1 where CITY_NAME='City B'
intersect
select PERSON_NO from table1 where CITY_NAME='City C'

Answer 2

Another way is to just join the table to itself a few times.

SELECT DISTINCT Person_No FROM mytable 
INNER JOIN mytable mt2 on (cityname='city b' and mt1.person_no=mt2.person_no)
INNER JOIN mytable mt3 on (cityname='city c' and mt1.person_no=mt3.person_no)
WHERE cityName='city a'

When first learning SQL most students don't realize it is completely legal to include the same table in a query multiple times. There are lots of problems that can be solved this way.

Answer 3

You can use multiple EXISTS() clauses:

   select a.PERSON_NO
   from personTable a
   where exists (select 1 from personTable where PERSON_NO = a.PERSON_NO
                 and CITY_NAME = 'City A')
   and exists (select 1 from personTable where PERSON_NO = a.PERSON_NO
                 and CITY_NAME = 'City B')
   and exists (select 1 from personTable where PERSON_NO = a.PERSON_NO
                 and CITY_NAME = 'City C')

Answer 4

I would do this with an aggregation and a HAVING . Assuming no repeats:

 SELECT person_no, count(*)
   FROM mytable
  WHERE city_name IN ('City A', 'City B', 'City C')
  GROUP BY person_no
 HAVING count(*) = 3

This has the advantage of hitting the table only once, which a join might not do.

Answer 5

Didn't actually test it, but this general idea should work:

SELECT DISTINCT PERSON_ID
FROM YOUR_TABLE T1
WHERE
    NOT EXISTS (
        SELECT CITY_NAME -- All cities.
        FROM YOUR_TABLE T2
        EXCEPT -- Or may be "MINUS", depending on your DBMS.
        SELECT CITY_NAME -- Cities the user worked in.
        FROM YOUR_TABLE T3
        WHERE T1.PERSON_NO = T3.PERSON_NO
    )

In plain English:

• For every distinct person, find the set difference between all cities and cities she worked in .
• If that difference is empty, the person has worked in all cities.

Answer 6

Try this

SELECT   PERSON_NO 
FROM     YOURTABLENAME
WHERE    CITY_NAME =  'City A'
         AND  CITY_NAME = 'City B'
         AND  CITY_NAME = 'City C'