Wierd proc_df behaviour

(Akash ) #1

Hi, i have a training set of size 1460x81. when i used proc_df, x size became 1460x83

my code:
train_cats(df_raw)
x,y,nas= proc_df(df_raw, ‘SalePrice’)
x.shape

the proc_df function:
same as structured.py

0 Likes

(Zachary Mueller) #2

Did your categorical have any NA’s? See if any column names now include _na. This is an autogenerated binary value that declares if the value was there or not. My guess is you have two of those columns

1 Like

(Akash ) #4

ohh i completely forgot about that. thanks a lot :smiley:

0 Likes