Export CDT Dekadal data about cdms.products

[lilyclements]

The ‘export_cdt’ function works the same for dekadal as for daily etc so there won’t be any parameter changes, just calling the export_cdt function directly.
However am happy to do this still, and adding an additional check the data type is dekadal. Can do the same with daily data?

[dannyparsons]

That's good they will work the same!

I think the difference will just be that for daily data the date column will be the date for each day whereas for dekadal data this column will need to be in the format: 1999011, 1999012, 1999013, 1999021 etc. which is like the number of each dekad.

In the way that our dekad function works currently this would need to be paste0(year, month, dekad) to get this format since dekad returns 1, 2, 3....

So I think we would just need to do this bit of pre-processing in the function. I think we keep it as one function with these options.

[lilyclements]

Sure. I assumed that the data in this function would have had the dekad function already applied.
So shall I change the “date” parameter to “type” (or whichever the parameter name is in the other _export function) and then it can read in a “date” variable for daily data, and the dekadal column for the dekad data.

[lilyclements]

Should it have a check in? If it works for any date format at the moment then a check might lose the functionality if, say, someone used pentad

[dannyparsons]

Yes, a type = c("daily", "dekad") argument would be good. The optional parameters date, year, month, dekad.
If type = "daily" then only date is required. The others are ignored. This is used for the output.
If type = "dekad" then we need date or year, month, dekad. We make a new column paste0(year, month, dekad) for use in the output. If only date is provided we can create year, month, dekad from it.

[dannyparsons]

I'm not sure if CDT accepts formats other than daily and dekad, which is what Climsoft exports to. If other formats are possible, we can always add those later.