Azure ML Logistic Regression Randomizing

new to the site!

I have two questions:

1) I am using logistic regression for a dataset but have too many 0 values versus the 1 values. This is causing my machine learning training/scoring/evaluating modules to not identify the 1 values (a lot of false positives). Can I randomly sample from the 0s to balance the dataset?

2) I have attempted to write code in Python to randomly select thvalues from my selected column of choice, and recode these as np.nan values.