How to convert columns of dataframe based on best-fit types of values in them?

Question

Following this answer (number 4) I am trying to use df.convert_dtypes()

Pandas version: 0.25.3

Reproducible:

import pandas as pd
import numpy as np

data = {
    "int": np.zeros((5, ), dtype=np.int),
    "float": np.zeros((5, ), dtype=np.float),
    "string": ["a", "b", "c", "ddd", "FFFFF"],
    "bool": [True, False, True, True, False]
}

df = pd.DataFrame(data)

print(df)
print(df.dtypes)

output so far:

   int  float string   bool
0    0    0.0      a   True
1    0    0.0      b  False
2    0    0.0      c   True
3    0    0.0    ddd   True
4    0    0.0  FFFFF  False
int         int32
float     float64
string     object
bool         bool
dtype: object

df.convert_dtypes()
print(df.dtypes)

I would expect the trivial types, rather than object, but getting

AttributeError: 'DataFrame' object has no attribute 'convert_dtypes'

What's the way to do this?

what version of pandas are you using? – Tomer Nov 18 '20 at 13:55 — Tomer, Nov 18 '20 at 13:55
Need pandas `1.+`, for me working nice. – jezrael Nov 18 '20 at 13:56 — jezrael, Nov 18 '20 at 13:56
well, looks like i am way behind wow – Gulzar Nov 18 '20 at 13:57 — Gulzar, Nov 18 '20 at 13:57

score 0 · Accepted Answer · answered Nov 18 '20 at 18:44

0

pip install --upgrade pandas

Thanks to @jezrael need version 1.+

answered Nov 18 '20 at 18:44

Gulzar

23,452
27
113
201

How to convert columns of dataframe based on best-fit types of values in them?

1 Answers1