Add functionality to create new column with different values based on matches

[JohannesWiesner]

https://stackoverflow.com/questions/68869434/create-an-pandas-column-if-a-string-from-a-list-matches-from-another-column

[JohannesWiesner]

For example, this would create new column "file_type" that has the values "image" and "motion" depending on whether the path in the filepath column contains ".nii" or ".txt":

# add info if files is image or motion files
pattern_dict = {".nii": "image", ".txt": "motion"}
df['file_type'] = df['filepath'].apply(
    lambda path: [val for key, val in pattern_dict.items() if key in path][0])

Would be nice if one could also make this work with regexes. The following code goes already in the right direction but it will only replace the found match (not the whole string) with the value:

# Create a new column based on pattern matching
pattern_dict = {r'^.nii$': 'image', r'^.txt$': 'motion'}
df['file_type'] = df['filepath'].replace(pattern_dict, regex=True)

[JohannesWiesner]

This would be the corresponding regex:

^(?=.abc).

https://chatgpt.com/c/673b627e-70b8-8011-a90d-649a949fca76

But then we could also simply use the above and check if substring is in string

[JohannesWiesner]

This would be the corresponding code:

pattern_dict = {"^(?=.*.nii).*":"image", "^(?=.*.txt).*": "motion"}
df['file_type'] = df['filepath'].replace(pattern_dict,regex=True)