4 ms·

After a few years of Clojure(Script), where immutability means you can use almost anything as keys, I have found many cases where my python code would have been

After a few years of Clojure(Script), where immutability means you can use almost anything as keys, I have found many cases where my python code would have been simpler and easier if I could use arbitrary objects (and therefore data structures) as keys. I mean, its not a huge deal, but it adds up.

orf 10y ago
> I have found many cases where my python code would have been simpler and easier if I could use arbitrary objects as keys You can use pretty much anything as a dictionary key
- dkersten 10y ago
  That's really not true[1]. You can only use hashable types, which excludes dicts, lists and sets. For example (tested on both python 2.7 and 3.4): >>> {{'a': 1}: 2} Traceback (most recent call last): File "<stdin>", line 1, in <module> TypeError: unhashable type: 'dict' >>> {[1,2,3]: 4} Traceback (most recent call last): File "<stdin>", line 1, in <module> TypeError: unhashable type: 'list' >>> {{1,2,3}: 4} Traceback (most recent call last): File "<stdin>", line 1, in <module> TypeError: unhashable type: 'set' Objects inheriting from object are hashable by default, but the hash is based on the objects instance such that each instance will return a unique value for that instance: >>> class A(object): ... def __init__(self): ... self.x = 1 ... >>> a = A() >>> b = A() >>> d = {a: 1, b: 2} >>> a.x = 10 >>> b.x = 10 >>> d[a] 1 >>> d[b] 2 While this makes sense for the instance stored in the dict (because mutability would otherwise mean that they key changes values), I don't think this is particularly useful. I rarely look up dicts by instance, but rather by value. That is, I would construct another object with the same attributes and look up by that. You can, of course, implement your own __hash__ to make it work, but you have to do it manually for any object you want it and most third-party objects won't have implemented it so you'd have to monkey patch them. Contrast that with Clojure, where all built-in data structures (maps, sets, lists, dicts etc) work as keys out of the box. [1] I guess its true in the sense that you can implement __hash__ to make it work. But its not particularly easy or idiomatic.

orf 10y ago

> I have found many cases where my python code would have been simpler and easier if I could use arbitrary objects as keys You can use pretty much anything as a dictionary key

dkersten 10y ago
That's really not true[1]. You can only use hashable types, which excludes dicts, lists and sets. For example (tested on both python 2.7 and 3.4): >>> {{'a': 1}: 2} Traceback (most recent call last): File "<stdin>", line 1, in <module> TypeError: unhashable type: 'dict' >>> {[1,2,3]: 4} Traceback (most recent call last): File "<stdin>", line 1, in <module> TypeError: unhashable type: 'list' >>> {{1,2,3}: 4} Traceback (most recent call last): File "<stdin>", line 1, in <module> TypeError: unhashable type: 'set' Objects inheriting from object are hashable by default, but the hash is based on the objects instance such that each instance will return a unique value for that instance: >>> class A(object): ... def __init__(self): ... self.x = 1 ... >>> a = A() >>> b = A() >>> d = {a: 1, b: 2} >>> a.x = 10 >>> b.x = 10 >>> d[a] 1 >>> d[b] 2 While this makes sense for the instance stored in the dict (because mutability would otherwise mean that they key changes values), I don't think this is particularly useful. I rarely look up dicts by instance, but rather by value. That is, I would construct another object with the same attributes and look up by that. You can, of course, implement your own __hash__ to make it work, but you have to do it manually for any object you want it and most third-party objects won't have implemented it so you'd have to monkey patch them. Contrast that with Clojure, where all built-in data structures (maps, sets, lists, dicts etc) work as keys out of the box. [1] I guess its true in the sense that you can implement __hash__ to make it work. But its not particularly easy or idiomatic.