Of course, you could instead take an "Easier to Ask Forgiveness Than Permission" approach:

try:
number = int(value)
except ValueError:
is_numeric = False
else:
is_numeric = True

This is all just to say "don't use those 3 string methods because it's too easy to mix them up and there are better options".

Convert to a number and catch the exception or use a regex to match the specific kind of number you need.

This week's tips will all be on regular expressions.

🧵 (6/6)

If you use a regular expression with "\d", it'll work the same way as isdecimal().

It'll match b, but not the other 2:

>>> import re
>>> re.fullmatch(r"\d+", b)
<re.Match object; span=(0, 1), match='５'>

Want to be more strict and match only "0" through "9"?

Use an explicit "[0-9]" character class:

>>> re.fullmatch(r"[0-9]+", b)

Or you could match negative numbers or numbers with a ".":

>>> re.fullmatch(r"-?[0-9]+\.?[0-9]*", "-5.0")
<re.Match object; span=(0, 4), match='-5.0'>

🧵 (5/6)

"\N{fullwidth digit five}" can be converted to a number:

>>> int(b)
5

The other two would raise a ValueError exception.

So, if you're going to use one of those 3 string methods, I'd use isdecimal().

The isdigit() and isnumeric() methods will match strings that aren't valid numbers (by Python's own definition of "number").

So... did you guess isdecimal() is the strictest of those 3?

No??

That's why I recommend against all of them.

🧵 (4/6)

>>> print(a, a.isnumeric(), a.isdecimal(), a.isdigit())
Ⅴ True False False
>>> print(b, b.isnumeric(), b.isdecimal(), b.isdigit())
５ True True True
>>> print(c, c.isnumeric(), c.isdecimal(), c.isdigit())
⓹ True False True

Roman numeral matches 1, double circle digit 2, and fullwidth digit 3!

Okay... so some obscure characters return True for some of those methods and False for others.

But why do we care?

Well, which can be converted to a number using the built-in int function?

🧵 (3/6)

All 3 methods return True for the first one only:

>>> print(a, a.isnumeric(), a.isdecimal(), a.isdigit())
5 True True True
>>> print(b, b.isnumeric(), b.isdecimal(), b.isdigit())
5.0 False False False
>>> print(c, c.isnumeric(), c.isdecimal(), c.isdigit())
-5 False False False

Huh.

Of the 3 symbols Ⅴ, ５, and ⓹, which will one will each of those match?

>>> a = "\N{roman numeral five}"
>>> b = "\N{fullwidth digit five}"
>>> c = "\N{double circled digit five}"
>>> print(a, b, c)
Ⅴ ５ ⓹

🧵 (2/6)

Or maybe you'll just reference the same variable(s) on both sides of an assignment while using tuple unpacking.

For example, here's an implementation of the Fibonacci sequence in Python:

def fibonacci():
a, b = 0, 1
while True:
a, b = b, a+b
yield a

Note that "a, b = b, a+b" assignment.

It's not quite variable swapping, but it relies on the same idea.

So while "swap 2 variables in one line" is likely pointless, the idea behind the trick does come in handy.

🧵 (2/2)