BNlearn R error “variable Variable1 must have at least two levels.”

Question

Trying to create a BN using BNlearn, but I keep getting an error;

Error in check.data(data, allowed.types = discrete.data.types) : variable Variable1 must ha         


        

Answer 1

bnlearn expects a data.frame : doesn't work with tibbles, So keep your data as a data.frame by omitting the line DataFull <- as_tibble(DataFull)

Example

library(tibble)
library (bnlearn)

d <- as_tibble(learning.test)
hc(d)

Error in check.data(x) : variable A must have at least two levels.

In particular, it is the line from bnlearn:::check.data

if (nlevels(x[, col]) < 2) 
      stop("variable ", col, " must have at least two levels.")

In a standard data.frame,learning.test[,"A"] returns a vector and so nlevels(learning.test[,"A"]) works as expected, however, by design, you cannot extract vectors like this from tibbles : d[,"A"]) is still a tbl_df and not a vector hence nlevels(d[,"A"]) doesn't work as expected, and returns zero.