Question

I’ve been asked to modify a report (which unfortunately was written horribly!! not by me!) to include a count of days. Please note the “Days” is not calculated using “StartDate” & “EndDate” below. The problem is, there are multiple rows per record (users want to see the detail for start & enddate), so my total for “Days” are counting for each row. How can I get the total 1 time without the total in column repeating?

This is what the data looks like right now:

ID          Description     startdate   enddate        Days
REA145681   Emergency       11/17/2011  11/19/2011      49
REA145681   Emergency       12/6/2011   12/9/2011       49
REA145681   Emergency       12/10/2011  12/14/2011      49

REA146425   Emergency       11/23/2011  12/8/2011       54
REA146425   Emergency       12/9/2011   12/12/2011      54

I need this:

ID          Description     startdate   enddate        Days
REA145681   Emergency       11/17/2011  11/19/2011      49
REA145681   Emergency       12/6/2011   12/9/2011       
REA145681   Emergency       12/10/2011  12/14/2011      

REA146425   Emergency       11/23/2011  12/8/2011       54
REA146425   Emergency       12/9/2011   12/12/2011     

Help please. This is how the users want to see the data.
Thanks in advance!

Liz

— Here is the query simplified:

    select id
,description
,startdate     -- users want to see all start dates and enddates
,enddate
,days = datediff(d,Isnull(actualstardate,anticipatedstartdate) ,actualenddate) 
    from table

Answer 1

As you didn’t provide the data of your tables I’ll operate over your result as if it was a table. This will result in what you’re looking for:

select *, 
  case row_number() over (partition by id order by id)
    when 1 then days
  end
from t

Edit:

Looks like you DID added some SQL code. This should be what you’re looking for:

select *, 
  case row_number() over (partition by id order by id)
    when 1 then
      datediff(d,Isnull(actualstardate,anticipatedstartdate) ,actualenddate)
  end
from t