Reputation: 13
I am trying to write a function that will subset a pandas dataframe based on the data type; essentially I am trying to split on numerical and non-numerical columns.
I have gotten as far as:
data.dtypes
This returns the types of each column, however I cannot find a list of all possible types. Currently it seems to me that all non-numerical columns are type 'object', is this the case?
Thanks
Upvotes: 1
Views: 839
Reputation: 25652
Pandas uses datetime64
, timedelta64
, int64
, float64
, bool
and object
(this includes strings and any other custom objects). You can use other types such as float32
, and pandas will try to maintain it, but some operations will implicitly cast to e.g., float64
.
Upvotes: 2