格式化本地人时出现KeyError [英] KeyError when formatting locals

问题描述

我面对的是KeyError我无法解释或理解. 我有一个笔记本，其中在单元格中定义了变量PREFIX:

I'm facing a KeyError I can't explain or understand. I have a notebook, in which I define a variable PREFIX in a cell:

PREFIX = "/home/mavax/Documents/info/notebook/log_study"

这只是指向包含日志的文件夹的路径，因此使用笔记本的人只需执行以下代码即可更改路径.

which is simply a path to a folder containing logs, so people using the notebook just need to change the path if they want to execute the code below.

然后，稍后(在下面有很多单元格)，我使用它，没有任何问题:

Then, later (quite a bunch of cells beneath), I use it, without any problem:

for basename in ["log_converted_full.txt", "log_converted_trimmed.txt"]:
    entries = load_log_for_insertion("%(PREFIX)s/datasets/logs/%(basename)s" % locals())
    pprint(entries)

然后我得到了期望的输出，这意味着找到了文件，并且正在打印日志中的(很长)输出.

I then get the output I expect, meaning files are found and the (very long) output from the logs is being printed.

我还有更多单元格描述我为解决此问题而实现的结构，当需要再次执行同一段代码时，我得到了KeyError:

I have some more cells describing the structure I implement for this problem, and when the time comes to execute again the same piece of code, I get the KeyError:

带来错误的代码:

def demo_synthetic_dig_dag(data_size):
    for basename in ["alert_converted_trimmed.txt"]:
        ###
        entries = load_log_for_insertion("%(PREFIX)s/datasets/logs/%(basename)s" % locals())[:data_size]
        g = AugmentedDigDag()
        g.build(entries)

        html(
            """
            <table>
                <tr><td>%s</td></tr>
            </table>
            """ % (
                synthetic_graph_to_html(g, 2, 0.03)
            )
        )

，然后在下一个单元格中:

and, in the next cell:

demo_synthetic_dig_dag(200)

Jupyter输出:

Jupyter output:

---------------------------------------------------------------------------
KeyError                                  Traceback (most recent call last)
<ipython-input-179-7c2a79d0afd6> in <module>()
----> 1 demo_synthetic_dig_dag_armen(200)

<ipython-input-178-d17f57de3c01> in demo_synthetic_dig_dag(data_size)
     18     for basename in ["log_converted_trimmed.txt"]:
     19         ###
---> 20         entries = load_log_for_insertion("%(PREFIX)s/datasets/logs/%(basename)s" % locals())[:data_size]
     21         g = AugmentedDigDag()
     22         g.build(entries)

KeyError: 'PREFIX'

我敢肯定这个错误很简单而且很愚蠢，但是如果有人可以睁开眼睛，我将非常感激！

I'm pretty sure the mistake is quite simple and plain stupid, but still, if someone could open my eyes, i'd be very thankful !

推荐答案

在函数之外，locals()与globals()相同，因此没有问题.

Outside a function, locals() is the same as globals(), so you have no issue.

但是，当放置在函数内部时，locals()不会以任何方式包含PREFIX(存储在globals()中，它包含该函数的本地名称.这就是为什么格式化它们会失败，正在尝试从locals()字典返回的字典中获取名为PREFIX的键.

When placed inside a function, though, locals() doesn't contain PREFIX in any way (it is stored in globals(), it contains the local names for that function. That's why formatting these fail, it's trying to get a key named PREFIX from the dictionary returned from the locals() dict.

为什么不使用.format进行格式化，而不是使用.format:

Instead of formatting with %, why not just use .format:

"{}/datasets/logs/{}s".format(PREFIX, basename)

或者，您可以将PREFIX引入本地范围，并为函数添加一个附加参数:

Alternatively, you could bring PREFIX in the local scope with an additional parameter to your function:

def demo_synthetic_dig_dag(data_size, PREFIX=PREFIX):

但是我看不出有什么好处. (是的，本地查找的性能会有小幅度提高，但我怀疑它会起到一定作用)

but I don't see much of an upside to that. (Yes, there is a small performance boost for local look-up but I doubt it would play a role)

这篇关于格式化本地人时出现KeyError的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！