Error in `[.data.frame`(x, , f) : undefined columns selected (in dismo)

[David Bennett]

Hi everyone,

Got a weird problem. I'm using dismo and R for a Maxent model and, when using the following code:

xm <- maxent(predictorsM1, pres_train, a=NULL, nbg=10000, args = "biasfile=biasfile", "replicates=10")

Get the following error:

Error in `[.data.frame`(x, , f) : undefined columns selected
In addition: Warning message:
In .local(x, p, ...) :
  7 (0.22%) of the presence points have NA predictor values

My search attempts on this group and google as a whole would suggest the problem lies with pres_train, but none of my efforts have 
worked. It is a spatial points data frame, but even changing it to a 2 column data frame doesn't work.

Would be everlastingly grateful for any help you could offer.

[Kyle Marsh]

I had the same error, but it in my situation it was because I ran the maxent model on a predictors stack with my "data" table for species locations. I copied the code from the documentation but didn't realize factors ="biome" didn't make sense for my data.  

maxent_model <- maxent(predictors, data, factors='biome', args=c('hinge=false', 'threshold=false'))

I didn't have any categorical variables called "biome".

Once I removed that from my code the error went away. 

For your situation:

Looking at your code it looks like you are not writing the args correctly. Try args = c( "biasfile=biasfile", "replicates=10").

Hopefully that helps,

Kyle