这是Python 3.5环境,我认为代码是自解释的,在这里,我希望这两个函数都能正常工作,但只有一个是正确的。
TLDR:
分配out = {**out, **answer}
会使out
变量包含右键,但在接下来的两个步骤中会丢失;虽然分配c = {**c,**a}
在测试功能中完美运行,但新密钥不会丢失。
有人可以解释一下我做错了吗?
def flatify_works(d, out, fhook=None):
for k, v in d.items():
if not isinstance(v, dict) and not isinstance(v, list):
out[k] = v
elif isinstance(v, dict):
flatify_works(v, out, fhook)
elif isinstance(v, list):
if fhook is None:
raise AssertionError("an array with more than 1 elment found.")
answer = fhook(k, v)
for k, v in answer.items():
out[k] = v
def flatify_doesnt_work(d, out, fhook=None):
for k, v in d.items():
if not isinstance(v, dict) and not isinstance(v, list):
out[k] = v
elif isinstance(v, dict):
flatify_doesnt_work(v, out, fhook)
elif isinstance(v, list):
if fhook is None:
raise AssertionError("an array with more than 1 elment found.")
answer = fhook(k, v)
out = {**out, **answer} # put a breakpoint here, and go 2 steps further
def hook(k, v):
l = [d["c"] for d in v]
return {"c": sum(l), "d": "blabla"}
def test_merge_dicts():
a = {"a": 1, "b": 2}
c = {"c": 3}
c = {**c, **a} # merging works perfectly here
print(c)
assert "a" in c and "b" in c and "c" in c # ok
def test_nested_works():
out = {}
flatify_works({"a": 1, "b": [{"c": 0.6, "d": 4}, {"c": 0.4, "d": 4}]}, out, hook)
print("working exemple: {}".format(str(out)))
def test_nested_doesnt_work():
out = {}
flatify_doesnt_work({"a": 1, "b": [{"c": 0.6, "d": 4}, {"c": 0.4, "d": 4}]}, out, hook)
print("not working exemple: {}".format(str(out)))
def main():
test_merge_dicts() # ok
test_nested_works() # ok
test_nested_doesnt_work() # why out = {**out, **answer} is not working as expected?
if __name__ == '__main__':
main()
答案 0 :(得分:2)
我认为区别在于,在第二个函数中,您重新分配out
变量,然后将其与传递给函数的dict分离:
out = {**out, **answer} # out is now a different object
由于函数是递归的,后续修改不会影响原始字典。
在第一个函数中,您只进行项目分配,它会按预期修改dict:
out[k] = v