Note: This question is certainly different from Does re in Python support word boundaries (/b). The alluded link seeks an answer for a very simple query for which a cursory glance on any tutorial in Python regular expression would have provided the explanation with examples. My question was using a word boundary around an OR expression and is far from trivial or to be reckoned as duplicate.
I was trying to build a palatable example to demonstrate regex word boundaries. Towards this, I wanted to show how the singular food items ordered by a diet-conscious person are changed for a guzzler and wrote the following program:
import re
items_lean = 'a masala dosa, an idli and a mango lassi'
pattern = r'{}'.format('an|a') # Use pattern as dynamic variable in regex
items_fat = re.sub(pattern, 'four', items_lean) # OOPS
print(items_fat)
pattern_fat = r'{}'.format('\ban\b|\ba\b') # Ensure a or an occurs as a word by itself
items_fat_proper = re.sub(pattern_fat, 'four', items_lean)
print(items_fat_proper)
I expected the following outputs corresponding to each print statement
four mfoursfourlfour dosfour, four idli fourd four mfourgo lfourssi
four masala dosa, four idli and four mango lassi
But, what I got was:
four mfoursfourlfour dosfour, four idli fourd four mfourgo lfourssi
a masala dosa, an idli and a mango lassi
Where should the \b factor be placed to get the desired output?
In order to satisfy the guzzlers you need to escape the \bs or use the raw input format i.e.
pattern_fat = r'\ban\b|\ba\b'
I've also removed the superfluous format which I suspect caused this confusion!