Unique Key Violation error while updating table

Question

Hi All,
I am having problem with UNIQUE CONSTRAINT. I am trying to update a table and getting the violation error. Here is the over view. We have a table called ActivityAttendee. ActivityAttendee has the following columns. The problem to debug is this table has
over 23 million records. How can I catch where my query is going wrong?
ActivityAttendeeID INT PRIMARY KEY IDENTITY(1,1)
,ActivityID INT NOT NULL (Foreign key to parent table Activity)
,AtendeeTypeCodeID INT NOT NULL
,ObjectID INT NOT NULL
,EmailAddress VARCHAR(255) NULL
UNIQUE KEY is on ActivityID,AtendeeTypeCodeID,ObjectID,EmailAddress
We have a requirement where we need to update the ObjectID. There is a new mapping where I dump that into a temp table #tempActivityMapping (intObjectID INT NOT NULL, intNewObjectID INT NULL)
The problem is ActivityAttendee table might already have the new ObjectID and the unique combination.
For example: ActivityAttendee Table have the following rows
1,1,1,1,NULL
2,1,1,2,NULL
3,1,1,4,'abc'
AND the temp table has 2,1
So essentially when I update in this scenario, It should ignore the second row because, if I try updating that there will be a violation of key as the first record has the exact value. When I ran my query on test data it worked fine. But for 23 million records,
its going wrong some where and I am unable to debug that. Here is my query
UPDATE AA
SET AA.ObjectID = TMP.NewObjectID
FROM dbo.ActivityAttendee AA
INNER JOIN #tmpActivityMapping TMP ON AA.ObjectID = TMP.ObjectID
WHERE TMP.NewObjectID IS NOT NULL
AND NOT EXISTS(SELECT 1
FROM dbo.ActivityAttendee AA1
WHERE AA1.ActivityID = AA.ActivityID
AND AA1.AttendeeTypeCodeID = AA.AttendeeTypeCodeID
AND AA1.ObjectID = TMP.NewObjectID
AND ISNULL(AA1.EmailAddress,'') = ISNULL(AA.EmailAddress,'')

Answer 1

>> I am having problem with UNIQUE CONSTRAINT. I am trying to update a table and getting the violation error. Here is the over view. We have a table called Activity_Attendee. <<
Your problem is schema design. Singular table names tell us there is only one of them the set. Activities are one kind of entity; Attendees are a totally different kind of entity; Attendees are a totally different kind of entity. Where are those tables? Then
they can have a relationship which will be a third table with REFERENCES to the other two. 
Your table is total garbage. Think about how absurd “attendee_type_code_id” is. You have never read a single thing about data modeling. An attribute can be “attendee_type”, “attendee_code” or “attendee_id”but not that horrible mess. I have used something like
this in one of my busk to demonstrate the wrong way to do RDBMS as a joke, but you did it for real. The postfix is called an attribute property in ISO-11179 standards. 
You also do not know that RDBMS is not OO. We have keys and not OIDs; but bad programmers use the IDENTITY table property (NOT a column!), By definition, it cannot be a key; let me say that again, by definition. 
>> ActivityAttendee has the following columns. The problem to debug is this table has over 23 million records [sic: rows are not records]<<
Where did you get “UNIQUE KEY” as syntax in SQL?? What math are you doing the attendee_id? That is the only reason to make it INTEGER. I will guess that you meant attendee_type and have not taken the time to create an abbreviation encoding it.
The term “patent/child” table is wrong! That was network databases, not RDBMS. We have referenced and referencing table. Totally different concept! 
CREATE TABLE Attendees
(attendee_id CHAR(10) NOT NULL PRIMARY KEY, 
 attendee_type INTEGER NOT NULL  --- bad design. 
    CHECK (attendee_type BETWEEN ?? AND ??), 
 email_address VARCHAR(255), 
CREATE TABLE Activities
(activity_id CHAR(10) NOT NULL PRIMARY KEY, 
Now the relationship table. I have to make a guess about the cardinally be 1:1, 1:m or n:m. 
CREATE TABLE Attendance_Roster
(attendee_id CHAR(10) NOT NULL --- UNIQUE??
   REFERENCES Attendees (attendee_id), 
 activity_id Activities CHAR(10) NOT NULL ---UNIQUE?? 
  REFERENCES Activities (activity_id)
 PRIMARY KEY (attendee_id, activity_id), --- wild guess! 
>> UNIQUE KEY is on activity_id, attendee_type_code_id_value_category, object_id, email_address <<
Aside from the incorrect “UNIQUE KEY” syntax, think about having things like an email_address in a key. This is what we SQL people call a non-key attribute. 
>> We have a requirement where we need to update the ObjectID. There is a new mapping where I dump that into a temp table #tempActivityMapping (intObjectID INTEGER NOT NULL, intNewObjectID INTEGER NULL) <<
Mapping?? We do not have that concept in RDBMS. Also putting meta data prefixes like “int_” is called a “tibble” and we SQL people laugh (or cry) when we see it. 
Then you have old proprietary Sybase UODATE .. FROM .. syntax. Google it; it is flawed and will fail. 
Please stop programming until you have a basic understanding of RDBMS versus OO and traditional file systems. Look at my credits; when I tell you, I think I have some authority. 
--CELKO-- Books in Celko Series for Morgan-Kaufmann Publishing: Analytics and OLAP in SQL / Data and Databases: Concepts in Practice Data / Measurements and Standards in SQL SQL for Smarties / SQL Programming Style / SQL Puzzles and Answers / Thinking
in Sets / Trees and Hierarchies in SQL