intermittent error in differentialAbundance?

[d-callan]

the following code:

myCollection <- getCollection(DiabImmune, '16S (V4) Species')

differentialAbundanceResults <- differentialAbundance(
  myCollection, 
  "breastfed_duration_days",
  groupA = function(x) {x < 300},
  groupB = function(x) {x >= 300},
  method = 'Maaslin2')  

sometimes produces this error:

Error in data.table::setnafill(df, fill = 0, cols = cols) :
  argument specifying columns received duplicate column(s)

and sometimes appears to not.. hard to diagnose until we manage to consistently repro.

[d-callan]

its unlikely an issue w differentialAbundance directly.. its confirmed (see #27 ) that the above collection does in fact have duplicate labels for some taxa, but should not have duplicate variables .. getCollection is meant to distinguish these cases, but perhaps is sometimes failing to?

perhaps a first step could be adding more detailed logging to the failing function, so that at least if we do see this again, well have a bit more information next time.

[d-callan]

i did manage to get my hands on a collection (using diabimmune genus again) w the following behavior, at one point:

> dupsIndex <- duplicated(getCollectionVariableNames(myCollection))
> getCollectionVariableNames(myCollection)[dupsIndex]
[1] "Coprobacter secundus"        "Eggerthella lenta"          
[3] "Flavonifractor plautii"      "unclassified Incertae Sedis"
[5] "unclassified Incertae Sedis" "unclassified Incertae Sedis"

but then requesting the collection again via getCollection only minutes later seemed to work properly..