mergeSequenceTables says what's wrong when stopping

[wbazant]

I've had the error that "At least two valid sequence tables, and no invalid objects, are expected." when running the function on contents of a folder. I didn't know which of the ~600 tables is wrong, and to find out that it was 553_featureTable.rds, I needed to do a bit of digging: find where the error comes from, open the R session with is.sequence.table , etc.

If the error message says what the problem is, it might save the user a bit of time.

Same rationale for duplicate sample names.

Also adds a check for the case of less than two tables passed in.

[benjjneb]

Small thing: Can you change "table" to "sequence table" in the error messages.

Second: What happens in the second error message if the tables are unnamed?

Third: There will be cases in which a large number of sample names (or maybe table names) are duplicated. This can lead to overlong and confusing error messages. The preferred behavior here would be to report a max number of duplicates, with an ellipsis at the end.

[wbazant]

Thank you, all good points!

Made another commit with better messages.

I've checked what happens when tables are unnamed: names() seems to return NULL, and paste0(NULL) returns an empty string.

Also stop() (or at least my version of it in the interactive R session) truncates long error messages, it doesn't have the ... but to me it looks kind of clear what happened, I wouldn't try to improve on this.

[benjjneb]

Look good except perhaps for https://github.com/benjjneb/dada2/pull/932/files#diff-4d02e5797bedf79c524bc1b3b7aeed18R314

I posted detailed comment there.

[benjjneb]

LGTM, thanks!