search

BiologicalRecordsCentre / sparta

Species Presence/Absence R Trends Analyses
http://biologicalrecordscentre.github.io/sparta/index.html
MIT License
21 stars 24 forks source link

formatOccDat() error and solution #125

Closed galinajonsson closed 5 years ago

galinajonsson commented 5 years ago

I came across the following error when running formatOccDat() (on windows this time)

  1. Error in vapply(indices, fun, .default) : values must be type 'logical', but FUN(X[[1380]]) result is type 'integer' I fixed it by changing fill=FALSE to fill=0 here: https://github.com/BiologicalRecordsCentre/sparta/blob/a8a3d0063e2d384af2332b02e417bad163170a6d/R/formatOccData.r#L151 As per this thread's advice: https://stackoverflow.com/questions/4835202/error-with-custom-aggregate-function-for-a-cast-call-in-r-reshape2

I'm happy to make a push request with the new code but thought I'd ask since I don't know whether it could affects other stuff (and frankly I don't fully understand the above thread).

AugustT commented 5 years ago

Your right, this looks like the sort of change that needs a little thought. Changing this to fill = 0 means that entries that were previously FALSE will now be 0. I dont think this is what we want to do as we want the result to be all Trues and Falses.

If you are debugging can you see what temp looks like just before this line is run? use str to look at its structure. There might be something wrong with that?

galinajonsson commented 5 years ago

You're absolutely right - although the above change allowed the function to run, the resulting df$spp_viswas full of 1s and 0s...

temp looks fine to me or what do you think Tom? Browse[2]> str(temp) 'data.frame': 43686 obs. of 3 variables: $ species_name: Factor w/ 54 levels "Aglais io","Aglais urticae",..: 46 46 17 17 17 46 40 39 32 7 ... $ visit : chr "HU31HU3716 25/08/1973" "HU44HU4741 25/06/1974" "HY20HY2509 8/1905" "HY20HY2509 9/1927" ... $ pres : logi TRUE TRUE TRUE TRUE TRUE TRUE ...

galinajonsson commented 5 years ago

I identified the issue! Some survey names contained NAs (e.g. "SiteA1997-07-NA") and the function ran fine when I removed these. Thanks for the help Tom.

AugustT commented 5 years ago

Oh interesting...I would not have thought that would cause problems..