用于删除数字中可变长度序列的正则表达式匹配

Question

I have a numpy array of integers that look like this X000Y000. X and Y can be of length 1, 2, or 3 and might contain 0s. I want to transform every element in the array to just X. I feel like a regex could be used for this but cannot figure out a good one to use, or how to apply that to a whole array.

Example: 14000010000 should become 140.

Answer 1

I assume X and Y can't begin with 0 . [1-9]\\d{0,2} matches a number from 1 to 3 digits that doesn't begin with 0 .

So the regexp to extract X and Y should be:

^([1-9]\d{0,2})000([1-9]\d{0,2})000$

Then you can use re.sub() to remove the zeroes between X and Y .

regex = re.compile(r'^([1-9]\d{0,2})000([1-9]\d{0,2})000$');
i = 14000010000
istr = str(i)
new_i = int(regex.sub(r'\1\2', istr)

You can map this over your numpy array

regex = re.compile(r'^([1-9]\d{0,2})000([1-9]\d{0,2})000$');
new_array = np.fromiter((int(regex.sub(r'\1\2', str(x)) for x in array), array.dtype)

See Most efficient way to map function over numpy array for various ways of mapping a function over a numpy array.