It's great article! Thanks for sharing. There the use case was to provide holidays with minimum date attributes, while here its more focus on try deriving all the possible date attributes.

To include holidays, there is CTE (HOLS) used which I personally would maintain as separate dimension along with country dimension and wouldn't add to this table which I stated above.

The reasoning behind is, to have an enterprise solution and I prefer normalized data sets, to maximize the scope of utilization and data maintenance.

Just my opinion and there is no right or wrong. So, nothing to fix in above article and it should work great for relavant appropriate use cases.