fix (cutData.R): don't overwrite existing "year" columns w/ yearseason

[jack-davison]

This PR fixes #404. cc @mooibroekd.

In short, when using cutData() w/ "yearseason" openair would overwrite existing year columns.

I've renamed the interim columns to be openair__year and openair__month, which are then deleted before the data is returned.

Using interim columns feels dodgy; a safer bet would be working on the vectors themselves I think. But this fix should work for the time being.

I note the function does leave a floating season column in the output which the user has not requested, but I don't want to remove it too in case the user already has a season column.

There may be broader improvements needed to how cutData() works in future - it can be very destructive to the user's own data if they're not careful.

devtools::load_all()
#> ℹ Loading openair

# create some date data for illustration
test <- data.frame(date = seq.Date(
  lubridate::date("2022-01-01"),
  lubridate::date("2022-12-31"),
  by = 1
))

# cut the data into years
test <- cutData(test, type = "year")

# check if we have any dates that are associated with 2023
test |>
  dplyr::filter(year == 2023)
#> [1] date year
#> <0 rows> (or 0-length row.names)

# now cut the data to yearseason
test <- cutData(test, type = "yearseason")

# perform the test again:
test |> filter(year == 2023)
#> [1] date       year       season     yearseason
#> <0 rows> (or 0-length row.names)

^{Created on 2025-01-22 with reprex v2.1.1}

[davidcarslaw]

I agree on cutData being too destructive and that isn't good. I think this is probably more of a concern with in-built types such as "year", "wd". However, even for any columns this is the case e.g. for "ws". I think a better solution would be to have a system that a new column is returned with a suffix e.g. ws_cut or ws_cat? That would make the outcome predictable for all variables - what do you think? It would require some different handling in openair functions that use cutData but that would be OK...