Views - SELECT from VIEW and SELECT from the query inside view definition returning different results

Question

Hi,
I am facing this weird issue. Any help would be appriciated.
I have view with the following definition:
CREATE VIEW [marketing].[OpenedMails]
AS
SELECT
ID_EmailAddress, 
ID_Date_Opened, 
ID_Contact, 
ID_MailSendJobs,
COUNT(ID_OpenedMails) AS OpenCount,
CASE
WHEN ROW_NUMBER() OVER (PARTITION BY CAST(ID_EmailAddress AS VARCHAR(10)) + ' ' + CAST(ID_MailSendJobs AS VARCHAR(10)) ORDER BY ID_Date_Opened) = 1 THEN 1 
ELSE 0 
END
AS UniqueOpenCount
FROM            
dbo.Fact_OpenedMails
where ID_Contact = 382340
GROUP BY ID_EmailAddress, ID_Date_Opened, ID_Contact, ID_MailSendJobs;
order by ID_MailSendJobs 
When I run the the select statement in the view definition I get combination of both 1 and 0 for the 'UniqueOpenCount' column.
But when I run the select from the view itself using the following query:
SELECT [ID_EmailAddress]
      ,[ID_Date_Opened]
      ,[ID_Contact]
      ,[ID_MailSendJobs]
      ,[OpenCount]
      ,[UniqueOpenCount]
  FROM [marketing].[OpenedMails]
I get equal amount of rows but only 0 values for the 'UniqueOpenCount' column which seems to be very weird to me. Why is this happening ? Can anyone help regarding how to solve this ??
Result from the select inside view definition:
Result from the select query directly from the view:
Thanks in advance.
Vivek Kamath

Answer 1

Please post DDL, so that people do not have to guess what the keys, constraints, Declarative Referential Integrity, data types, etc. in your schema are. Learn how to follow ISO-11179 data element naming conventions and formatting rules. You failed. Temporal
data should use ISO-8601 formats – you failed again. Code should be in Standard SQL AS much AS possible and not local dialect. 
This is minimal polite behavior on SQL forums. 
Things like “ID_Date_Opened” are wrong. The affixes “id” and “date” are called attribute properties in ISO-11179 and data modeling. Since a date is a unit of measurement on a temporal scale, it cannot be an identifier by definition. My guess is this would be
“open_date” if it were done right. And the only display format in ANSI/ISO Standard SQL is ISO-8601 (yyyy-mm-dd)
An email address of -1?? Email addresses are VARCHAR(256), not numeric. There is no reason to cast them AS string!
Your vague “mail_send_jobs” is plural, but columns hold scalars and cannot be plural by definition. The partition cause can hold a list of columns: 
WHEN ROW_NUMBER() 
     OVER (PARTITION BY email_address, mail_send_job 
               ORDER BY open_date)
     = 1 
THEN 1 ELSE 0 END 
This still makes no sense, but the syntax is better. You do not understand how ROW_NUMBER() works. 
Would you like to try again with proper Netiquette? 
--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