Python使用正则表达式在pandas列中提取字符串的一部分

Question

I've got a pandas df column called 'Raw' for which the format is inconsistent. The strings it contains look like that:

'(1T XXX, Europe)'
'(2T YYYY, Latin America)'
'(3T ZZ/ZZZZ, Europe)'
'(4T XXX XXX, Africa)'

The only thing consistent in the strings in 'Raw' is that they start with a digit, includes a comma in the middle followed by a whitespace, and they contain parentheses as well.

Now, I'd like to create two extra columns (Model and Region) in my dataframe:

• 'Model' would contain the beginning of the string, ie everything between the first parenthesis and the comma
• 'Region' would contain the end of the string, ie everything between the whitespace after the comma and the final parenthesis

How do I do that using regex?

Answer 1

Since there's only one comma, and everything is between parentheses, in your case, use .str.split() instead, after slicing appropriately:

model_region = df.Raw.str[1:-1].str.split(', ', expand = True)

But if you insist:

model_region = df.Raw.str.extract('\((.*), (.*)\)', expand = True)

Then

df['Model'] = model_region[0]
df['Region'] = model_region[1]

Answer 2

Try this : \\(([^,]*), ([^)]*)\\)

See : https://regex101.com/r/fCetWg/1

Answer 3

import re

s = '(3T ZZ/ZZZZ, Europe)'
m=re.search(r'\((.*), (.*)\)',s)
print(m.groups())

Answer 4

Model=re.findall(r"(?<=\().+(?=\,)",s)
Region=re.findall(r"(?<=\, ).+(?=\))",s)

The first regex checks for opening bracket "(" in front of the model and closing ",". The second regex checks for any string between "," and ")".

Answer 5

string_list = ['(1T XXX, Europe)',
'(2T YYYY, Latin America)',
'(3T ZZ/ZZZZ, Europe)',
'(4T XXX XXX, Africa)']
df = pd.DataFrame(string_list)
df = df[0].str.extract("\(([^,]*), ([^)]*)\)", expand=False)

Answer 6

If the comma is a reliable separator of your string parts, then you do not need regexp. If df is your dataframe:

df['Model'] = [x.split(',')[0].replace('(', '') for x in df['Raw']]
df['Region'] = [x.split(',')[1].replace(')', '') for x in df['Raw']]

if you want to use regexp is would look something like:

s = '(1T XXX, Europe)'
m = re.match('\(([\w\s]+),([\w\s]+)\)', s)
model = m.group(1)
region = m.group(2)