Database null row caused by "deleting duplicate rows with one statement"

-- Duplicate null columns can be detected.
Delete   From   [  Jcxs0201  ]   Where    User_id   In  (  Select   User_id U From   [  Jcxs0201  ]   Group   By  User_id    Having   Count (ID) >  1  )  And  ID  Not   In ( Select   Min (ID) From   [  Jcxs0201  ]  Where   User_id   In ( Select   User_id   From    [  Jcxs0201  ]   Group   By   User_id    Having   Count (ID) >  1 ))

 

-- Duplicate null Columns cannot be detected.
Delete   From   [  Jcxs0201  ]   Where    User_id   In  (  Select    User_id U From   [  Jcxs0201  ]    Group   By  User_id    Having   Count (ID) >  1  )  And  ID  Not   In ( Select   Min (ID) From    [  Jcxs0201  ]   Group   By   User_id    Having   Count (ID) >  1 )

 

PS: later repeated tests/2012 10 18

Found in (null ,~) Null is invalid, which is equivalent to in (~)

So here is a problem: If the column contains a null value in the group by column

It can only be described by a non-null value, such as the ID of the modified row or the row number obtained by the row_number () function.

 

Analysis of the preceding statement errors:

It is described in two parts: user_id and ID.

I. Check user_id first

1. The effects of Count (ID) and count (user_id) after group by user_id are different.

The following statement indicates that three rows of user_id are null, and count (null) seems invalid.

2. Make the userid in (null, 1257646) and user_id in (1257646) statements consistent, and the previous null will be ignored.

In fact, all results are user_id in (1257646 ).

Can it be understood that, in SQL Server, null is not a value, but a default status?

2. Check the group ID.

The statement containing grou by user_id does not know whether the select ID can be used.

If not, null cannot be processed. Therefore, the row_number () function must be used to add a row number to it.

This is my understanding. Please point out any errors