CODEWITHSUNDEEP
pythonpandastypes
lara_toff
lara_toff

Reputation: 442

Change datatype of columns in Pandas Dataframe depending on the original data type of the column

I have a large (200 columns) dataframe that has int64 and float64 columns. Without typing out all 200 column names, is it possible to convert all of the int64 to int32, and all of the float64 to float32?

Upvotes: 0

Views: 598

Answers (2)

cfort
cfort

Reputation: 2776

Select all a particular dtype columns:

int64_cols = df.select_dtypes(['int34']).columns.tolist()

Change those columns to int32:

df[int64_cols] = df[int64_cols].astype('int32')

Upvotes: 0

Jeremy
Jeremy

Reputation: 46

You only have to create the dictionary that will be passed in the method .astype() in pandas. Let's assume that data is your pandas dataframe of 200 columns.

dtype_dict = data.dtypes.replace({'int64': 'int32', 'float64': 'float32'}).to_dict()
data = data.astype(dtype_dict)

Upvotes: 3

Related Questions