Fix: Type instability in eqClassById during full_join in addEmptyRC

[ChangqingW]

Fixes a BiocParallel error (specifically a dplyr::full_join incompatible types error) that occurs when a eqClassCount have only one row.

The column eqClassById is normally a list-column, but when there is only one row in the tibble, it gets simplified in to a atomic integer vector, making it incompatible to full_join with the minEquiRC tibble:

Tracing addEmptyRC(eqClassCount, annotations) on entry
Called from: eval(expr, p) 
 Browse[1]> eqClassCount
# A tibble: 1 × 4
  eqClassById GENEID                 nobs rcWidth
        <int> <chr>                 <int>   <int>
1          -2 ENSMUSG00000025290.17     1     514 

 Error: BiocParallel errors
  1 remote errors, element index: 2
  2 unevaluated and other errors
  first remote error:
Error in full_join(eqClassCount, minEquiRC, by = c("eqClassById", "GENEID", : Can't join `x$eqClassById` with `y$eqClassById` due to incompatible types. 

Added a line to force it into a list column.

[ChangqingW]

I had a look and turned out it was probably because one of my test cases has only 1 read. I guess this is a very extreme edge case that doesn't really matter...

[cying111]

Hi @ChangqingW , thanks for the updates! Good to see extreme cases where our code might break.

I will close this PR for now.

Feel free to re-open it if you find it necessary.

Thank you