Random sampling many times and run through glm model

I have a large dataset from which i need to take a random sample many
times ( say N=50) and run it through the same glm() - logistic regression
model everytime ( 50 times )  and capture the chi-square p-values ( Pr >
ChiSq ) of the variables for each run and output  average p-value of the
variables that went into the model. 

I have done this in SAS. LIke to know how i can do this in R. Any help is
appreciated.