Daily count of employees

Question

Hi! I'm new, am hoping for some help with a particular query. I need to count the number of employees that were active at any given location on any given day. For each active employee, there would be one row for each day, as I need to be able to run the query "as of" a certain point in time. We are only interested in dates beginning 1-1-2011 going forward.
An example of a result set would be:
EMPNO DEPTNO LOCNO EMPSTATUS JOBNO HIRE_DATE TERM_DATE CHANGE_EFF_DATE COUNT_DATE COUNT
111 111 101 Active 111 21-may-2007 01-sep-2012 0
111 112 101 Active 123 21-may-2007 01-may-2012 01-sep-2012 1
222 202 102 Terminated 333 5-june-2010 1-sep-2012 01-sep-2012 1
333 303 301 Terminated 444 2-sep-2010 6-aug-2011 01-sep-2012 0
444 100 202 Active 222 1-jun-2010 01-sep-2012 1
Total active employees on 1-sep-2012 = 3
111 112 101 Active 123 21-may-2007 01-may-2012 02-sep-2012 1
222 202 102 Terminated 333 5-june-2010 1-sep-2012 02-sep-2012 0
444 100 202 Active 222 1-jun-2010 02-sep-2012 1
Total active employees on 2-sep-2012 = 2
In the above example, the "TERM_DATE" is actually the employee's last day of employment, so I would count them if the term date = count date. Also, Employee # 111 has changed job numbers, so is not counted twice on the same day. Employees terminated before the count date would not be counted. We would filter by job number, location, department in our BI tool.
I hope this explains what I am trying to do. I will try and clarify if need be. Thanks to anyone for any help!

Answer 1

Hi,
Welcome to the forum!
truesoprano wrote:
Here is some additional information. The database version is 11.2.0.2, Great! That's important.
The other things mentioned in the forum FAQ {message:id=9360002} are important, too; in particular, CREATE TABLE and INSERT statements for your sample data, and using \ tags to make your output readable.
and I have looked at serveral things in trying to find out how to do this (google, SQL book, techonthenet, ittoolbox). I've found a lot of "employee counts" and "date increments" but nothing that seems like it will quite fit my scenario.
The source data comes from JD Edwards employee master and a custom history table, I have that filtered.
Would someone mind taking a look at this and point me in the right direction? Thank you so much.Do you want the output to contain 1 row for every day, starting with Septemebr 1, and continuing through today?  If you don't already have a table containing those dates, then use a sub-query, like dates_of_interest, below.  Outer-join that to your actual table, to find which employees were working on which days.
I think you'll want something like this:WITH     dates_of_interest     AS
     SELECT     DATE '2012-08-31' + LEVEL     AS a_date
     FROM     dual
     CONNECT BY     LEVEL <= SYSDATE - DATE '2012-08-31'
SELECT     d.a_date
,     COUNT (DISTINCT x.empno)     AS a_date_cnt
FROM          dates_of_interest d
LEFT OUTER JOIN     table_x     x ON x.cnt     > 0     -- COUNT is not a good column name
                    AND d.a_date     BETWEEN x.hire_date
                                   AND     x.term_date
GROUP BY d.a_date
ORDER BY d.a_date
Obviously, I can't test it until you post CREATE TABLE and INSERT statements for your sample data.