feat: accept NA in countElements; see #61

[sgibb]

This PR allows NAs in subtractElements and addElements by handling NAs in countElements (problem described in #61 )

subtractElements("H2O", NA) returns "H2O". Or should it return NA?

Unfortunately it introduce a strange behaviour for containsElements:

containsElements("H2O", NA)
# [1] TRUE

IMHO containsElements("H2O", NA) should return FALSE or NA.

@jorainer, @michaelwitting any suggestion for the expected behaviour?

[jorainer]

IMO the following would make sense:

containsElements("H2O", NA) should return NA
substractElement("H2O", NA) returns NA_character_
containsElements("H2O", "") should return FALSE (same as e.g. containsElements("H2O", "Z"))

with the current PR you're also reporting TRUE for containsElements("H2O", "Z")... thus, it always returns TRUE even if the provided element does not exist...

[jorainer]

See comment - basically I would suggest to change the behavior to return NA for countElements("H2O", NA) as well as for subtractElement or addElement where a NA is provided.

[sgibb]

containsElements("H2O", NA) should return NA substractElement("H2O", NA) returns NA_character_

done!

containsElements("H2O", "") should return FALSE (same as e.g. containsElements("H2O", "Z"))
with the current PR you're also reporting TRUE for containsElements("H2O", "Z")... thus, it always returns TRUE even if the provided element does not exist...

This wasn't introduced by this PR but is a bug that was/is already present. I add an issue #63 .

IMHO we could close #61 with this PR and fix #63 in another one.

[jorainer]

Thanks @sgibb for the PR! Can you please also bump the version and add an entry to the NEWS file? apart from that all OK from me to merge.