Dataiku is telling me that is_numeric_dtype does not exist.

Iceberg

Hi,

I created a Python recipe to standarize columns

# -*- coding: utf-8 -*-
import dataiku
import pandas as pd, numpy as np
from dataiku import pandasutils as pdu

# apply the z-score method in Pandas using the .mean() and .std() methods
def z_score(df):
# copy the dataframe
df_std = df.copy()
# apply the z-score method
for column in df_std.columns:
if is_numeric_dtype(df_std[column]):
df_std[column] = (df_std[column] - df_std[column].mean()) / df_std[column].std()

return df_std

# Read recipe inputs
spam_prepared = dataiku.Dataset("spam_prepared")
spam_prepared_df = spam_prepared.get_dataframe()

spam_standarized = z_score(spam_prepared_df)

# Write recipe outputs
spam_standarized = dataiku.Dataset("spam_standarized")
spam_standarized.write_with_schema(spam_standarized_df)

But Dataiku is telling me that is_numeric_dtype does not exist. What am I doing wrong?

Thanks

AgatheG

Hi Iceberg,

is_numeric_dtype is a method from the module pandas.api.types. In your code snippet, you should replace your occurrences of is_numeric_dtype with pd.api.types.is_numeric_dtype.

Best regards,

Agathe