Restricting Uniques from Showing

问题

I currently have this SQL query and this output.

SELECT b.HotelNo,g.guestName,b.dateFrom,b.dateToFrom Booking b, Guest g
FROM Booking b, Guest g
WHERE b.guestNo = g.guestNo
GROUP BY b.hotelNo,b.dateFrom,b.dateTo,g.guestName;

I am getting this output: Output

How do I remove all the rows that are not duplicates of first two columns?

Output would be (In words Get all the hotel numbers guest names and dates of stay of all guest only if they have stayed in the same hotel more than once):

1234 John Doe 2017-02-01 00:00:00.000 2017-02-28 00:00:00.000
1234 John Doe 2017-03-16 00:00:00.000 2017-03-21 00:00:00.000

EDIT:
I got what I wanted but I when I add b.dateFrom,b.dateTo into my select statement. I get an error saying: Column 'Booking.dateFrom' is invalid in the select list because it is not contained in either an aggregate function or the GROUP BY clause.

This is the SQL query that gets this error:

SELECT b.hotelNo,g.guestName,b.dateFrom,b.dateTo 
FROM Booking b, Guest g 
WHERE b.guestNo = g.guestNo
GROUP BY b.hotelNo,g.guestName 
HAVING COUNT(*) > 1;

If I add those two column names in (which I need) I get the error above. How do I add them in without getting an error?

回答1:

If you want only those who have stayed in a hotel more than once, use either a JOIN or an EXISTS clause to check whether the count is greater than two.

Using EXISTS:

SELECT DISTINCT b.hotelNo,g.guestName,b.dateFrom,b.dateTo 
FROM Booking AS b
JOIN Guest AS g 
    ON b.guestNo = g.guestNo
WHERE EXISTS 
(
    SELECT 1
    FROM Booking
    WHERE hotelNo = b.hotelNo
    AND guestNo = b.guestNo
    GROUP BY hotelNo, guestNo
    HAVING COUNT(*) > 1
);

Using a JOIN (but same principle):

SELECT DISTINCT b.hotelNo,g.guestName,b.dateFrom,b.dateTo 
FROM Booking AS b
JOIN Guest AS g 
    ON b.guestNo = g.guestNo
JOIN
(
    SELECT hotelNo, guestNo
    FROM Booking
    GROUP BY hotelNo, guestNo
    HAVING COUNT(*) > 1
) AS j
    ON j.hotelNo = b.hotelNo
    AND j.guestNo = b.guestNo;

NOTE: I'm not sure that you need DISTINCT or a GROUP BY on your columns unless you actually have duplicates in your tables (and if you do, you probably should get rid of those duplicates).

回答2:

If you don't mind picking only one of the "dateFrom"s (first or last depending on the order clause) then you can assign a row number to each row, partitioning by the first two columns. Then only pick the first rows of each group (RN=1) which has the same first two column

WITH CTE AS
( SELECT ROW_NUMBER()   
          OVER ( PARTITION BY b.HotelNo,g.guestName 
            ORDER BY b.dateFrom) AS RN,
         b.HotelNo,g.guestName,b.dateFrom,b.dateTo
  FROM Booking b, Guest g
  WHERE b.guestNo = g.guestNo
  GROUP BY b.hotelNo,b.dateFrom,b.dateTo,g.guestName;
)
SELECT HotelNo,guestName,dateFrom,dateToFrom 
WHERE RN=1

You can read more HERE

来源：https://stackoverflow.com/questions/43083226/restricting-uniques-from-showing

标签

sql

sql-server

duplicates