[FIXED] Using a python function to replace characters in specific column then throws TypeError: cannot convert the series to <class 'float'>

Issue

I want to use clean_currency method below to strip out commas and dollar signs from Savings ($) column and put into a new column called 'Savings_Clean'.

def clean_currency(curr):
    return float(curr.replace(",", "").replace("$", ""))

clean_currency("$60,000")#The output is proof that function is working

Output: 60000.0

How to I clean Savings ($) column, because when I put clean_currency(df.Savings) I get the following error: TypeError: cannot convert the series to <class 'float'>

Solution

To make it work with ordinary strings and also the pd.Series that contain strings, an if guard can be used:

def clean_currency(curr):
    if isinstance(curr, str):
        return float(curr.replace(",", "").replace("$", ""))
    return curr.str.replace(",", "").str.replace("$", "").astype(float)

Now it will look the argument passed curr and if it is a string, cast with float, otherwise assume it is a series and use astype. (note that we use str accessor for replacing the series).

>>> clean_currency("$60,000")
60000.0

>>> clean_currency(pd.Series(["$60,000", "$120,000", "$1,000,000"]))
0      60000.0
1     120000.0
2    1000000.0
dtype: float64

Answered By - Mustafa Aydın