如何更改Pandas中除第一列之外的所有列的列类型？

Question

I have a 6,000 column table that is loaded into a pandas DataFrame. The first column is an ID, the rest are numeric variables. All the columns are currently strings and I need to convert all but the first column to integer.

Many of the functions I've found don't allow passing a list of column names or drop the first column entirely.

Answer 1

You can do:

df.astype({col: int for col in df.columns[1:]})

Answer 2

An easy trick when you want to perform an operation on all columns but a few is to set the columns to ignore as index:

ignore = ['col1']

df = (df.set_index(ignore, append=True)
        .astype(float)
        .reset_index(ignore)
       )

This should work with any operation even if it doesn't support specifying on which columns to work.

Example input:

df = pd.DataFrame({'col1': list('ABC'),
                   'col2': list('123'),
                   'col3': list('456'),
                  })

output:

>>> df.dtypes
col1     object
col2    float64
col3    float64
dtype: object

Answer 3

Try something like:

df.loc[:, df.columns != 'ID'].astype(int)

Answer 4

Some code that could be used for general cases where you want to convert dtypes

# select columns that need to be converted
cols = df.select_dtypes(include=['float64']).columns.to_list()
cols = ... # here exclude certain columns in cols e.g. the first col
df = df.astype({col:int for col in cols})

You can select str columns and exclude the first column in your case. The idea is basically the same.