Python：将lambda与startswith一起使用

Question

I need to writing my dataframe to csv, and some of the series start with "+-= ", so I need to remove them first.

I tried to test by using a string:

test="+++++-= I love Mercedes-Benz"
while True:
    if test.startswith('+') or test.startswith('-') or test.startswith('=') or test.startswith(' '):
        test=test[1:]
        continue

    else:
        print(test)
        break

Output looks perfect:

I love Mercedes-Benz.

Now when I want to do the same while using lambda in my dataframe:

import pandas as pd

col_names =  ['A', 'B', 'C']
my_df  = pd.DataFrame(columns = col_names)
my_df.loc[len(my_df)] = ["++++-= I love Mercedes-Benz", 4, "Love this"]
my_df.loc[len(my_df)] = ["=Looks so good!", 2, "5-year-old"]
my_df

my_df["A"]=my_df["A"].map(lambda x: x[1:] if x.startswith('=') else x)
print(my_df["A"])

I am not sure how to put 4 startswith "-","=","+"," " together and loop them until they meet the first alphabet or character(sometimes it might be in Japanese or Chinese.)

expected final my_df:

         A                    B          C
0   I love Mercedes-Benz      4       Love this
1   Looks so good!            2       5-year-old

Answer 1

You can use str.lstrip in order to remove these leading characters:

my_df.A.str.lstrip('+-=')

0     I love Mercedes-Benz
1           Looks so good!
Name: A, dtype: object

Answer 2

The function startswith accepts a tuple of prefixes:

while test.startswith(('+','-','=',' ')):
    test=test[1:]

But you can't put that in a lambda. But then, you don't need a lambda: just write the function and pass its name to map .

Answer 3

One way to achieve it could be

old = ""
while old != my_df["A"]:
    old = my_df["A"]
    my_df["A"]=my_df["A"].map(lambda x: x[1:] if any(x.startswith(char) for char in "-=+ ") else x)

But I'd like to warn you about the strip() method for strings:

>>> test="+++++-= I love Mercedes-Benz"
>>> test.strip("+-=")
' I love Mercedes-Benz'

So your data extraction can become simpler:

my_df["A"].str=my_df["A"].str.strip("+=- ")

Just be careful because strip will remove the characters from both sides of the string. lstrip instead can do the job only on the left side.

Answer 4

As a lover of regex and possibly convoluted solutions, I will add this solution as well:

import re

my_df["A"]=my_df["A"].map(lambda x: re.sub('^[*-=\s]*', '', x))

the regex reads:
^ from the beginning
[] items in this group
\\s any whitespace
* zero or more
so this will match (and replace with nothing) all the characters from the beginning of the string that are in the square brackets