how do i filter columns with data_type= object-CodePudding

encoder=LabelEncoder()
categorical_features=df.columns.tolist()
for col in categorical_features:
    df[col]=encoder.fit_transform(df[col])
df.head(20)

**i want categorical_features to take columns with datatype=object

CodePudding user response：

Try this

for col in categorical_features:
    if df.dtypes[col] == "object":
        print('object type')
    else:
        print(f'something else {df.dtypes[col]}')

you can modify your code like

categorical_features=df.columns.tolist()
for col in categorical_features:
    if df.dtypes[col] == "object"
       df[col]=encoder.fit_transform(df[col])

CodePudding user response：

to select all object type columns in pandas you can use:

categorical_features = df.select_dtypes(include=['object'])

The select_dtypes() was released in Pandas v 0.14.1