filling NAN and converting to int pandas

Question

I have a dataframe of integers. Preview (starts from 3 due to first 3 rows removal):

The original data in the 'pixel1' column is int, but the NAN there forced it to float.

I tried to fix it with:

X_train.fillna(method='ffill', inplace=True)
X_train = X_train.astype(int)
print(X_train.head())

that results in:

can I get the datatype of the value the fillna is using?
is there a better way to do so? (better = to skip the astype step, as the data is int originally - I planted the NAN in the file and that caused the int to float unwanted data conversion...)

`fillna` uses the same data type the column already has. What do you mean "a better way", is there something wrong with your current approach? — jdehesa, May 23 '18 at 09:07

score 2 · Accepted Answer · answered May 23 '18 at 09:06

2

I suggest use ffill with bfill for back filling if possible some NaNs in first row:

X_train = X_train.ffill().bfill().astype(int)

If not:

X_train = X_train.ffill().astype(int)

answered May 23 '18 at 09:06

jezrael

1 Answers1