RatesCurvesCsvLoader: 'Date' vs. 'Day count' vs. 'Label column'

When loading rates curves from CSV the following documentantion page show the desired structure of the files: RatesCurvesCsvLoader (OpenGamma Strata)

I see different columns regarding date and day calculations, such as, the Date and Label (e.g., 1M, 2M) columns in the values file and the Day Count (e.g. ACT/360) column in the settings file. I would like to know how the data in these columns interacts.

My primary question is: If I provide discount factors (as has been recommended to me here) with these file, I have to provide a Date value and a Label value in my file. For instance:

2019-02-20,USD-Disc,2019-03-21,0.949,1M

But couldn’t Strata on its own calculate the Date column (that is, 2019-03-21) if I provide a Holiday Calendar and a Day Count? From the label, Strata knows that it is the 1M node. So based on that info, Strata should be able to add one month to the value date (that is, 2019-02-20). I am just wondering why I have to provide the date (2019-03-21), if all the other info is already there.

A followup question is: How do I typically calculate the Date column when having a, say, 3M forward rate? I know that the Day Count is important, but also a Holiday Calendar?

Thanks

That class is useful when loading curves that have been calibrated elsewhere. In those situations, the date will have been calculated when performing the calibration. The "1M"is just a description used when the results need explaining to a human, so can’t be used for calculating a date.

The system will use the day count and the date provided to calculate the double x-value associated with the curve y-value you provide.