Re Regular expression operations, remove periods?-CodePudding

I'm working with a function I made to split this sample line below to remove the standalone numerical values (123), however it's also removing the trailing numbers which I need. I also can't figure out how to remove the "0.0"

ABC/0.0/123/TT1/1TT//

cleaned_data = []
def split_lines(lines, delimiter, remove = '[0-9] $'):
  for line in lines:
    tokens = line.split(delimiter)
    tokens = [re.sub(remove, "", token) for token in tokens]
    clean_list = list(filter(lambda e:e.strip(), tokens))
    cleaned_data.append(clean_list)
    print(clean_list)
split_lines(lines, "/")

What's coming out now is below, notice the 0. and "TT" that's missing the trailing 1.

[ABC], [0.], [TT], [1TT]

CodePudding user response：

Do you really need regular expressions? This job is much simpler if you just use str.split() and try to convert the resulting values to float:

def split_lines_remove_numeric(lines, delimiter):
    for line in lines:
        clean_list = []
        for item in line.split(delimiter):
            if not item: continue # Skip this item if it's empty
            try: 
                # Convert to float
                float(item)
            except ValueError: # Enter this block if conversion threw an error
                clean_list.append(item)
        print(clean_list)

Then, calling this function removes the values you want:

>>> split_lines_remove_numeric(["ABC/0.0/123/TT1/1TT//"], "/")
['ABC', 'TT1', '1TT']

CodePudding user response：

Try including the start of line anchor (^) as well.

cleaned_data = []
def split_lines(lines, delimiter, remove = '^[0-9.] $'):
  for line in lines:
    tokens = line.split(delimiter)
    tokens = [re.sub(remove, "", token) for token in tokens]
    clean_list = list(filter(lambda e:e.strip(), tokens))
    cleaned_data.append(clean_list)
    print(clean_list)
split_lines(lines, "/")

I simply changed the default value of the remove parameter to '^[0-9.] $' which only matches if the entire search string is numbers (or a period).