should plot_KDE() support numeric vectors and 1-column data frames?

[mcol]

The docs for 'plot_KDE()' state:

#' @param data [data.frame] or [RLum.Results-class] object (required):
#' for data.frame: two columns: De (values[,1]) and De error (values[,2]).
#' For plotting multiple data sets, these must be provided as
#' list (e.g. list(dataset1, dataset2)).

However, the code accepts also vectors and doesn't actively check that a data.frame contains two columns:

 if(is(data[[i]], "RLum.Results") == FALSE &
         is(data[[i]], "data.frame") == FALSE &
         is.numeric(data[[i]]) == FALSE) {
      .throw_error("Input data must be one of 'data.frame', ",
                   "'RLum.Results' or 'numeric'")

So the following errors can be generated:

plot_KDE(df[, 1])
# Error in data[[i]][, 1:2] : incorrect number of dimensions
plot_KDE(df[, 1, drop = FALSE])
# Error in `[.data.frame`(data[[i]], , 1:2) : undefined columns selected

Do we want to support these uses and update the documentation, or should we start rejecting them?

[RLumSK]

Please remove the vector, we only allow what is documented. No idea why it ended up there.

[mcol]

We've rediscussed this issue considering that:

• the current code is already dealing with numeric vectors and 1-column data frames: in such cases, it creates a 2-column data frame by appending a column of NAs
• however, currently this leads to all rows being removed, as the code calls 'na.exclude()` without checking if the second column was artificially added by us
• calc_Statistics() throws a warning: "All errors are NA or zero! Automatically set to 10^-9!"

We've reached the following conclusions:

• we will support numeric vectors and 1-column data frames and update the documentation accordingly
• instead of adding a column of NAs, we will add a column of very small values, such as 10^(-9) so that the behaviour of calc_Statistics() will not change while silencing the warning

[mcol]

Fixed by #194.