Change data type of a specific column of a pandas dataframe

Question

I want to sort a dataframe with many columns by a specific column, but first I need to change type from object to int. How to change the data type of this specific column while keeping the original column positions?

Answer 1

df['colname'] = df['colname'].astype(int) works when changing from float values to int atleast.

Answer 2

You can use reindex by sorted column by sort_values, cast to int by astype:

df = pd.DataFrame({'A':[1,2,3],
                   'B':[4,5,6],
                   'colname':['7','3','9'],
                   'D':[1,3,5],
                   'E':[5,3,6],
                   'F':[7,4,3]})

print (df)
   A  B  D  E  F colname
0  1  4  1  5  7       7
1  2  5  3  3  4       3
2  3  6  5  6  3       9

print (df.colname.astype(int).sort_values())
1    3
0    7
2    9
Name: colname, dtype: int32

print (df.reindex(df.colname.astype(int).sort_values().index))
   A  B  D  E  F colname
1  2  5  3  3  4       3
0  1  4  1  5  7       7
2  3  6  5  6  3       9

print (df.reindex(df.colname.astype(int).sort_values().index).reset_index(drop=True))
   A  B  D  E  F colname
0  2  5  3  3  4       3
1  1  4  1  5  7       7
2  3  6  5  6  3       9

If first solution does not works because None or bad data use to_numeric:

df = pd.DataFrame({'A':[1,2,3],
                   'B':[4,5,6],
                   'colname':['7','3','None'],
                   'D':[1,3,5],
                   'E':[5,3,6],
                   'F':[7,4,3]})

print (df)
   A  B  D  E  F colname
0  1  4  1  5  7       7
1  2  5  3  3  4       3
2  3  6  5  6  3    None

print (pd.to_numeric(df.colname, errors='coerce').sort_values())
1    3.0
0    7.0
2    NaN
Name: colname, dtype: float64

Answer 3

I have tried following:

df['column']=df.column.astype('int64')

and it worked for me.

Answer 4

To simply change one column, here is what you can do: df.column_name.apply(int)

you can replace int with the desired datatype you want e.g (np.int64), str, category.

For multiple datatype changes, I would recommend the following:

df = pd.read_csv(data, dtype={'Col_A': str,'Col_B':int64})