Objective: Create Staff Turnover Report from scratch - sql

Question

I have a requirement to create staff turnover report that should display following information according to parameterized period:
current staff
starters
leavers
manager
contracted site
Currently, I have a few ideas how to get a report. Either create slowly changing dimension or copy snapshots of employees table for each period or track starter and leavers seperately for each payrun. I have tried slowly changing dimension and it did fine job on new starters; however, it failed on leavers. I am not sure if tracking starters and leaver for each payruns would work.
Created a table below as a start of solution
I am relatively new in MS SQL 2008 business. Please adise. The main goal is to identify managers and contracted sites with highest staff turnover.
Notes
-Currently, I am not using any specific techonlogy and looking for solution.
-I do intend to build and design a brand new database for the sake of the report. By database I mean a few collection of tables that sit on CRM database (SQL 2008). By Collection of tables I mean "dimension" table for managers, "dimension" table of contracted sites, and a table of staff. So currently, I think that my solution should have at least 3 tables; however, my skills are not that good to solve the puzzle.
I think that have tables above would let me to write a SQL query that could compare different periods and get required results.
payno depcod Idd Imported
12568 EDE322001 12568EDE322001A 31 December 2011
12568 EDE322001 12568EDE322001B 31 December 2011
16822 EDE322001 16822EDE322001A 31 December 2011
17694 EDE322001 17694EDE322001A 31 December 2011
12568 EDE322001 12568EDE322001A 04 January 2012
12568 EDE322001 12568EDE322001B 04 January 2012
16822 EDE322001 16822EDE322001A 04 January 2012
17694 EDE322001 17694EDE322001A 04 January 2012
12568 EDE322001 12568EDE322001A 31 January 2012
12568 EDE322001 12568EDE322001B 31 January 2012
16822 EDE322001 16822EDE322001A 31 January 2012
17694 EDE322001 17694EDE322001A 31 January 2012
17661 EDE322001 17661EDE322001A 31 January 2012
12568 EDE322001 12568EDE322001A 01 February 2012
12568 EDE322001 12568EDE322001B 01 February 2012
16822 EDE322001 16822EDE322001A 01 February 2012
17906 EDE322001 17906EDE322001A 01 February 2012
17907 EDE322001 17907EDE322001A 01 February 2012
12568 EDE322001 12568EDE322001A 29 February 2012
12568 EDE322001 12568EDE322001B 29 February 2012
17907 EDE322001 17907EDE322001A 29 February 2012
Table above is table of snapshots. Snapshot date is displayed in Imported Column. ( Applogoes for messy table; I could not figure out how to make a table) Now i need to find a away how to compare one date to another (ideally in the loop) to figure out the difference between 1st date and following date.
For instance, 17661 EDE322001 17661EDE322001A 31 January 2012 is a new staff and 16822 EDE322001 16822EDE322001A 01 February 2012 is a leaver.
Many thanks

Answer 1

I may be looking at this too simplistically given all the talk about slowly changing dimensions, etc, but what's wrong with a plain old bit of SQL? I'm going to assume you have an Employee table with Manager, ContractedSite, StartDate and EndDate fields.
You want the following:
Current staff. Is this current at the start of the period or at the end of the period or active at some time during the period?
Starters. People who have a StartDate greater than or equal to the start of the period but less than or equal to the end of the period.
Leavers. People who have an EndDate greater than or equal to the start of the period but less than or equal to the end of the period.
Turnover. Which brings us back to what you mean by "Current staff". Turnover is essentially the difference between the current staff at the start of the period and the current staff at the end of the period, expressed either as a raw number or a percentage. So current staff at the start of the period are people with a StartDate less than the period start and either no EndDate or an EndDate greater than or equal to the start of the period. Similarly with the current staff at the end of the period.
So you have some specific cases to extract from your data:
SELECT Manager, ContractedSite,
SUM(CASE WHEN StartDate < #PeriodStart AND ((EndDate IS NULL) OR (EndDate >= #PeriodStart)) THEN 1 END) AS OpeningStaff,
SUM(CASE WHEN StartDate >= #PeriodStart AND StartDate <= #PeriodEnd THEN 1 END) AS Starters,
SUM(CASE WHEN EndDate >= #PeriodStart AND EndDate <= #PeriodEnd) THEN 1 END) AS Leavers,
SUM(CASE WHEN StartDate <= #PeriodEnd AND ((EndDate IS NULL) OR (EndDate > #PeriodEnd)) THEN 1 END) AS ClosingStaff
FROM Employee
WHERE (StartDate <= #PeriodEnd) AND ((EndDate IS NULL) OR (EndDate >= #PeriodStart))
GROUP BY Manager, ContractedSite
ORDER BY Manager, ContractedSite
Now you just need to calculate the difference between OpeningStaff and ClosingStaff in your report to get the turnover and you're done.