[Python] Python running probit model: len() of unsized object

I am trying to run a super simple python script where I pull data from a sql query and then extract the appropriate variables to run a probit on. nl and the variable names listed in cols are all long values.

import statsmodels.api as smf
import numpy as np
import pandas as pd
df = spark.sql("select * from knn_df")
cols=['constant','clnt_tenure_0_1', 'clnt_tenure_2_5', 'clnt_tenure_6_9',
'clnt_tenure_10_19', 'clnt_tenure_20p']
X = df[cols]
X = np.array(X)
y = df[['nl']]
y= np.array(y)
probit_model = smf.Probit(y,X)
result=probit_model.fit()
print(result.summary2())


When running this code I get the error "len() of unsized object". Any help would be appreciated.

Continue reading...