如何将我的函数应用于DataFrame列？

Question

I've defined the following function:

def clearString(myString):
    forbidden = r'/\:*?"<>|'
    for character in forbidden:
        if character in myString:
           myString = myString.replace(character,'')
    return myString

To remove unwanted characters in file names. I have a data frame with book titles in a column and I'm trying to apply the function to all the strings inplace, to clear them, but have been unable to, I keep getting the DataFrame back with untouched data.

I've already tried the apply function, both in the column alone and the entire DataFrame, and none of that yields a positive result, be it assigning the DataFrame back to it self, as in:

df = df.apply(clearString)
#Or even
df = clearString(df)

Or even defining a new one:

df_new = df.apply(clearString)
#Or even
df_new = clearString(df)

Is there something wrong with my function maybe, like not properly handling DataFrames or something?

Answer 1

apply isn't working because, by default, it applies the given function to each column (and not to each element). In the given examples, clearString would receive a Series argument, not a str .

To apply a function to all the elements of a DataFrame, one can use the applymap method ( docs ).

Examples:

# if you wanna replace the old dataframe
df = df.applymap(clearString)

# if you wanna keep the old dataframe
new_df = df.applymap(clearString)

Answer 2

您可以使用地图，甚至可以应用和地图组合。

Answer 3

If you want to modify a single column you can try these approaches:

df = pd.DataFrame({"Title": ["one ", "two", "three", "four"]})
def clean(title):
    return title.upper()
df["Title"] = df["Title"].apply(lambda x: clean(x))
# OR 
df["Modified_Title"] = df["Title"].apply(lambda x: clean(x))
# OR 
df["Modified_Title1"] = df.apply(lambda x: clean(x["Title"]), axis=1)
# OR 
new_df = df.applymap(lambda x: clean(x))