Data structure - horizontal vs vertical

[kyle_fredette]

I am currently building a solution in Excel that will allow users to enter information on a template and have SAS import the template for calculations. The template will include start and end dates for a given condition. In this case, the condition refers to the department that employees work within. Since employees change departments, we want to track data based on which department they completed the work for.

Currently, I am looking at this in two different ways (see charts below). Table one has the employees enter changes by including new rows of data. Table two has the employees enter changes by including new columns of data. Does anyone have experience with which of these data structures works best for longevity and flexibility? I want the best long term solution.

Thanks for your time!

Kyle

Vertical

Employee	Department	Start date	End date
John	Finance	1/1/2015	3/31/2015
John	Marketing	3/31/2015	12/31/2015
Mary	Operations	1/1/2015	12/31/2015
Susan	Sales	1/1/2015	12/31/2015

Horizontal

Employee	Department	Start date	End date	Team	Start date	End date
John	Finance	1/1/2015	3/31/2015	Marketing	3/31/2015	12/31/2015
Mary	Operations	1/1/2015	12/31/2015
Susan	Sales	1/1/2015	12/31/2015

[Steelers_In_DC]

I think vertical is the way to go, add a date field for when the record was added and changed and you can sort by that if necessary, horizontal can get ugly depending on how long this is in use.

[Reeza · Accepted solution]

Vertical, it is a common data structure, and you can easily determine which is the latest/first via sorting by date. You can either leave the end date missing for current positions or set it to something like Dec 31 9999.

FYI This is called a Type 2 Slowly Changing Dimension (SCD)

Slowly changing dimension - Wikipedia, the free encyclopedia

[AnnaBrown]

Also, LinusH wrote an article series on SCDs that may be helpful.