I mean this would remove False and None from a list though.
Also 0 and empty strings
And empty lists, tuples, dictionaries, sets, and strings
results = list(filter(None, results))results = [result for result in results if result != None]Should be
is not None(:You’re right, though IIRC there’s no functional difference when comparing to None (other than speed).
Yes there is. One invokes
__ne__on the left hand side, the other performs an identity comparison.
Meanwhile in Java land:
Map map = new Map()Map is an interface though, you’d have to use HashMap 😅
I don’t visit often…
Should have used memory safe rust.
nailed it
People used to argue that Python was incredibly readable. Then I started seeing shit like this.
I’d say this is pretty readable
It jams far too much on one line. Break it up. It’s a mistake I see a lot on Python.
but if you do it as a for loop it would be slower
Unless you’re looping over more than a million elements, that’s a poor excuse.
or you’re doing this a million times
I think that Python has a bit of a ‘Microsoft Word’ thing on the go. You know how your own docs are completely editable and print fine, but everyone else’s are a complete fucking disaster and pressing a single key will screw up the formatting of the whole document? Your own Python code is full of sensible idioms and pragmatic naming conventions, but everyone else’s was plainly written while on mushrooms.
I don’t know python, but it’s clear what that line does
Python has great list comprehension. Too bad it’s incomprehensible to humans.
But seriously whoever writes those cool oneliners is a shitty programmer. Life is not code golf, fuck off with your smart ass.
What would be the alternative? (assuming that you want to do the loop yourself)
new_results = [] for result in results: if result: new_results.append(result) results = new_resultsor else
for result in results: if not result: results.remove(result)which doesn’t do the exact same thing.
Honestly, this list comprehension is much faster to read and quite easy to understand.
I think we could rename the “result” variable “x” or “res” and it would be less confusing though.
[ L I S T C O M P R E H E N T I O N ]
Ruby has a method for this :)
[1] pry(main)> vars = ["one", "two", nil, "three"] => ["one", "two", nil, "three"] [2] pry(main)> vars.compact => ["one", "two", "three"]In Ruby,
0and""is truthy, so this would be a little different than the Python interpretation. You could filter with #select, but you’d typically write your code around this instead.In Ruby,
0and""is truthy,What the fuck?
Lua is the same. Only
falseandnilare “falsey”.I remember I fixed a bug in a mod for Minetest that assumed
0was false and caused an infinite loop.And people bash Javascript as if it was the devil when thinks like this exist on other languages.
Yup :) Everything in Ruby inherits Object, too. It’s a really neat language, and when you get accustomed to it, you might wonder why some other languages aren’t written like it.
For the 0 value being truthy, consider that Ruby is a dynamic language. Imagine if you asked a user how many motorcycles they own. If they answer, you’ll have an Integer. If they don’t, you’ll have nil, from NilClass. 0 is just as valid of an answer as 2, so you can do something like this:
raise NoResponse unless motorcycles save_motorcycles(motorcycles)deleted by creator
Python also has about 9000 alternatives that are better than this.
Allowing anything other than variables, number literals, and ‘:’ inside list indices was a mistake.
I’ve written this more times than I can remember 😹 who needs filter anyway? Gotta use up all this ram.
It does! it takes a list (or other iterator) and filters out all values that are cast to boolean True. The same could be archived with
results = list(filter(bool, results))It would filter out values that cast to False, no?
Like None, 0, “”, …
i wasn’t sure how to phrase it, it keeps all values that cast to True, and discards all that cast to False
Python’s disdain for the industry standard is wild. Every other language made in the last 20 years has proper filtering that doesn’t require collecting the results back into a list after filtering like Java (granted it’s even more verbose in Java but that’s a low bar).
If Python had modern lambdas and filter was written in an inclusion or parametric polymorphic way, then you could write:
new_results = results.filter(x -> x)Many languages have shorthands to refer to variables too, so it wouldn’t be impossible to see:
new_results = results.filter(_)Of course in actual Python you’d instead see:
new_results = list(filter(lambda x: x, results))which is arguably worse than
new_results = [x for x in results if x]hey if you can think of a better way
True.
Checks out
Looks like something I’d write while high
