how to convert data with float, int and strings to just strings and float

Question

I have a csv with data in different formats. there are float, int and strings.

How can I modify just the int into float?

I have not tried much, just with this:

df=df.astype(dtype=np.float64, copy=True, errors='raise')

but cant manage to make it work, because there are also strings

How can I solve this?

you can convert string to float if they're actually number i.e. a string '3.2' can be converted, a string like 'hello' can't be converted. You may want to convert the string into its ASCII code and convert it with ` float()` but it's some kind of process to include — NicoCaldo
– NicoCaldo, Commented Jul 7, 2022 at 12:29

Matteo Buffagni · Accepted Answer · 2022-07-07 12:30:49Z

1

The columns of a pandas DataFrame (or a Series) are homogeneously of type. You can inspect this with dtype (or DataFrame.dtypes)

You can select_dtypes first and and finally convert to float using df.astype

tmp = df.select_dtypes(include=['Int64'])
df[tmp.columns]= tmp.astype('float64')

answered Jul 7, 2022 at 12:30

Matteo Buffagni

3323 silver badges10 bronze badges

Sign up to request clarification or add additional context in comments.

1 Answer 1