简短但SQL查询（T-SQL，SQL Server）

Question

Suppose I have 2 tables, each tables has N columns. There are NO duplicate rows in table1

And now we want to know what datasets in table2 (including duplicates) are also contained in table1 .

I tried

select * from table1
intersect
select * from table2

But this only gives me unique rows that are in both tables. But I don't want unique rows, are want to see all rows in table2 that are in table1 ...

Keep in mind!! I cannot do

select *
from table1 a, table b
where a.table1col = b.table2col

...because I don't know the number of columns of the tables at runtime.

Sure I could do something with dynamic SQL and iterate over the column numbers but I'm asking this precisely because it seems too simple a query for that kind of stuff..

Example:

create table table1 (table1col int)
create table table2 (table2col int)

insert into table1 values (8)
insert into table1 values (7)

insert into table2 values (1)
insert into table2 values (8)
insert into table2 values (7)
insert into table2 values (7)
insert into table2 values (2)
insert into table2 values (9)

I want my query then to return:

8
7
7

Answer 1

If the amount of columns is not know, you will have to resort to a value computed over a row to make a match.

One such function is CHECKSUM .

Returns the checksum value computed over a row of a table, or over a list of expressions. CHECKSUM is intended for use in building hash indices.

SQL Statement

SELECT  tm.*
FROM    (
          SELECT  CS = CHECKSUM(*)
          FROM    Table2
        ) tm          
        INNER JOIN (          
          SELECT  CS = CHECKSUM(*)
          FROM    Table2          
          INTERSECT
          SELECT  CHECKSUM(*)          
          FROM    Table1
        ) ti ON ti.CS = tm.CS

Note that CHECKSUM might introduce collisions. You will have to test for that before doing any operation on your data.

Edit

In case you are using SQL Server 2005 , you might make this a bit more robust by throwing in HASH_BYTES .

The downside of HASH_BYTES is that you need to specify the columns on which you want to operate but for all the columns you do known up-front, you could use this to prevent collisions.

Answer 2

EXCEPT vs INTERSECT - link

EXCEPT returns any distinct values from the left query that are not also found on the right query.

INTERSECT returns any distinct values that are returned by both the query on the left and right sides of the INTERSECT operand.

Maybe EXCEPT can solve your problem