Any issue for a Date Dimension without day?

[computermike]

I have some data but it is aggregated at the Month level.  Would I have any issues created a Date table with the following columns  DateID(int), Date(date).. but only for the first i.e. 1/1/2017, 2/1/2017 etc.,  Month(string) "Jan","Feb", Year(int), MonthNUm(int)

I know I have to define as Date Table.  Any other considerations?  Will this work?

Thanks,

Mike  

[MattAllington · Accepted solution]

The is no issue with this. Just be aware that inbuilt time intelligence will not work - you will need to write custom time intelligence. 

Here are some of my articles FYI. 

http://exceleratorbi.com.au/power-pivot-calendar-tables/

http://exceleratorbi.com.au/dax-time-intelligence-beginners/

[computermike]

Yes, I was hoping to use time intelligence.  I guess I will create a table with all dates.  I thought I might be able to save space by skipping individual days.  Not advanced enough to write my own time intelligence functions.

I also decided I will not need some columns.  My new schema will be...    Date(date) i.e. 1/1/2017, 1/2/2017,     Month(string) i.e. "Jan", "Feb",    Year(number) i.e. 2016,2017.  I will be using Month and Year as Labels on my report.  Thought about making Year a string too, but don't think it ios that big of deal.  I will sort Month using Date.

Thanks,

Mike

[MattAllington]

---

@computermike wrote:

Yes, I was hoping to use time intelligence. 

You can still use time intelligence with a non-day level calendar, just not the inbuilt time intelligence. Custom time intelligence is easy once you learn how. Read my article on the topic posted earlier. I explain how it works and you can go from there. If you can put a filter in a table using the drop down arrows to filter the table to show the dates you need, then you can write custom time intelligence.

I guess I will create a table with all dates.  I thought I might be able to save space by skipping individual days. 

The space saving on a dimension table would be insignificant.