[R-package] fix warnings in demos

[jameslamb]

Similar to the changes made to examples in the R package's API docs in #4568, this PR fixes warnings in demos related to the deprecation warnings added as part of #4226.

For example, if you run Rscript R-package/demo/basic_walkthrough.R on current master, you'll see the following on stdout.

Warning message:
In lgb.train(data = dtrain2, num_leaves = 4L, learning_rate = 1, :
lgb.train: Found the following passed through '...': num_leaves, learning_rate, nthread, objective. These will be used, but in future releases of lightgbm, this warning will become an error. Add these to 'params' instead. See ?lgb.train for documentation on how to call this function.

This PR fixes them and converts examples to the new pattern we want to encourage as of v3.3.0 (#4310), passing all parameters through params.

Notes for Reviewers

By release 4.0.0, the goal is to have these demos converted to R package vignettes (#1944). However, until then I think it's at least worth updating these to the new recommended pattern, in case users reference the demos as examples.

[StrikerRUS]

Please fix duplicated min_data param here:

LightGBM/R-package/demo/categorical_features_rules.R

Lines 86 to 88 in ee5636f

	, min_data = 1L
	, learning_rate = 0.1
	, min_data = 0L

And update multiclass.R demo in the following two places:

LightGBM/R-package/demo/multiclass.R

Lines 27 to 28 in ee5636f

	, min_data = 1L
	, learning_rate = 1.0

LightGBM/R-package/demo/multiclass.R

Lines 42 to 47 in ee5636f

	, min_data = 1L
	, learning_rate = 1.0
	, early_stopping_rounds = 10L
	, objective = "multiclass"
	, metric = "multi_error"
	, num_class = 3L

[StrikerRUS]

Is it possible to pass keyword arguments before positional ones in R?

Suggested change

	params = params
	params

[jameslamb]

yes, R doesn't care about that difference. R's argument parsing will first match all values passed as keyword args, then pass remaining positional arguments to remaining unmatched parameters.

f <- function(x1, x2,  x3) {
    print(paste0("x1: ", x1, " | x2: ", x2, " | x3: ", x3))
}

f("a", "b", "c")
# [1] "x1: a | x2: b | x3: c"

f(x2 = "a", "b", "c")
# [1] "x1: b | x2: a | x3: c"

f("b", x3 = "a", "c")
# [1] "x1: b | x2: c | x3: a"

I'm fine with the suggestion, but just wanted to show R isn't opinionated about that the way Python is.

[StrikerRUS]

Oh, interesting! Thanks for the explanation and sorry for the false alarm.

[jameslamb]

Please fix duplicated min_data param here...and update multiclass.R

Fixed in 3838d33, thanks.

[StrikerRUS]

Nice, thank you! Just one minor suggestion below:

[github-actions]

This pull request has been automatically locked since there has not been any recent activity since it was closed. To start a new related discussion, open a new issue at https://github.com/microsoft/LightGBM/issues including a reference to this.