import re k = 2 s = '1110001110110' print("*CP refers to Capture Group\n") print(" ", f'{s} => Input Value\n') # ----------------------- print("Using lambda with:") print(" ", re.sub("1"*k, lambda m: m.group(0) + '0', s), "=> No *CP") print(" ", re.sub(f'({"1"*k})', lambda m: m.group(1) + '0', s), "=> Unnamed *CP\n") # ----------------------- print("Using \g with:") print(" ", re.sub("1"*k, '\g<0>0', s), "=> No *CP") print(" ", re.sub(f'({"1"*k})', '\g<1>0', s), "=> Unnamed *CP") print(" ", re.sub(f'(?P{"1"*k})', '\g<1>0', s), "=> Named *CG\n") # ----------------------- print("Using \g with:") print(" ", re.sub(f'(?P{"1"*k})', '\g0', s), "=> Named *CG") ''' Review the notes below to better understand the differences with each line above. ************ A. Numeric Backreference Example: ******** The 2nd parameter for re.sub(pattern, repl, string) can accept a string value with references to capture groups using something known as numeric backreferences (\1, \2, \n, ...). r'\1-' will append '-' to capture group 1. re.sub('(abc)', r'\1-', 'abcdef') # abc-def ---- ************ B. Ambiguous Backreference Example: ******** However, combining a numeric backreference with a numeric string literal, like '0' or '1', is ambiguous for the parser. Consider if in the the previous example, we wanted to append '0' instead of '-'. r"\10" would be ambiguous in that it could mean: - r'\1' + '0' - ...or it is the 10th capture group Consequently, the following will raise an error: ---- re.sub('(abc)', r'\10', 'abcdef') Throws invalid group reference exception. ---- One way to address this ambiguity is to use a lambda function that receives the matched object (m) which can be used to extract the value from group by index position. m.group(0) is equivalent to r'\1'. -------- Lambda example: ---- lambda m: m.group(0) + '0' ---- ************ C. Using Symbolic Group \g: ******** A symbolic group can refer to capture groups using the following format: \g => where '\g<1>' and '\1' are equivalent. -------- Unnamed Capture Group with \g: ---- re.sub(f'({"1"*k})', '\g<1>0', s) ---- A symbolic group name can be defined in the pattern by placing (?Pcapture_pattern) within the capture parenthesis as seen in the pattern below: f'(?P{"1"*k})' Named capture groups can be referenced using the the following format: \g => where '\g' and '\1' are equivalent. The group name can then be referenced in the replacement string as follows: -------- Using Unnamed Capture Group with \g: ---- re.sub(f'({"1"*k})', '\g<1>0', s) ---- ------------------- Useful References: ------------------- https://docs.python.org/3.8/library/re.html#index-17 https://docs.python.org/3.8/library/re.html#index-37 '''

Regex with Symbolic <named> Groups