如何替换所有字符串'？ 用 0 浮点数？

Question

I am cleaning a data frame currently and am running into issues because all of them are a mix of int and str, but I am trying to convert all of them to floats. The data frame is all numbers as well as some entries with '?'strings that I am trying to replace with '0' floats. How should I go about doing so?

    # Load the data from the file
    df = pd.read_csv('processed.state.csv')
    df.apply(pd.to_numeric) 

Yields an error: Unable to parse string "?" at position 165

Answer 1

df = pd.DataFrame([1,23,'1','2', "?"])
df.replace('?', 0).apply(pd.to_numeric)

A more generic solution to replace non-numbers to 0 will be

def fun(x):
  try:
    return float(x)
  except ValueError:
    return 0

df = pd.DataFrame({'c1': [1,23,'1','2', "?"], 'c2': [1,23,'abc','2', "?"]})

df.applymap(fun)

Answer 2

You can create your own function:

def to_float(item):
    try:
        return float(item)
    except ValueError:
        return 0

And apply that to the DataFrame instead.

Answer 3

You can use pandas.DataFrame.replace :

df = pd.read_csv('processed.state.csv' encoding = 'utf-8')
df.replace('?', 0)
df.apply(pd.to_numeric) 

Answer 4

df['col'] = df['col'].map(lambda x: 0.0 if x == '?' else x).astype(np.float64)