将Python中的文本网格转换为嵌套字典

Question

将Python中的文本网格转换为嵌套字典

3

想象一下一个文本文件中的网格，就像这样：

以下是将此内容转换为Python字典的最佳方法： ```python { "key1": "value1", "key2": "value2", "key3": "value3" } ``` 请注意，键和值可以根据需要更改。

{
  'A': {'A': 0, 'B':3, 'C':6},
  'B': {'A': 1, 'B':0, 'C':7},
  'B': {'A': 2, 'B':5, 'C':0}
}

所以我可以使用以下方式访问单元格:

matrix['A']['B'] # 3

我目前有一些非常粗糙的代码（请不要太苛刻地评价我）：

matrix = {}
f = open(filepath, 'r')
lines = f.readlines()
keys = lines[0].split()

for key in keys:
    matrix[key] = {}

for line in lines[1:]:
    chars = line.split()
    key_a = chars[0]
    for i, c in enumerate(chars[1:]):
        key_b = keys[i-1]
        matrix[key_a][key_b] = int(c)

print matrix

# Outputs {'A': {'A': 1, 'C': 0, 'B': 2}, 'C': {'A': 7, 'C': 6, 'B': 0}, 'B': {'A': 0, 'C': 3, 'B': 5}}

虽然这种方法没有错，但是我已经很长时间没有接触Python了，有更好的方法吗？也许嵌套字典并不是最好的方法？

更新：

不幸的是，我需要使用纯Python实现，因此无法使用外部库（相信我，我很想用）。
将样例代码从伪代码更新为实际代码。惭愧地低下头。

- Pete Hamilton

你能发一下你已经有的代码吗？ - Snakes and Coffee

1

完成，对眼睛造成不适表示歉意。 - Pete Hamilton

“虽然这样做并没有错...但有更好的方法吗？” - 你的问题最好发布在http://codereview.stackexchange.com上。 - Robᵩ

谢谢，我之前没看到过这个！以后有类似的问题会去那里问。 - Pete Hamilton

1个回答

网页内容由stack overflow 提供, 点击上面的

可以查看英文原文，
原文链接

- unutbu · Accepted Answer

您的代码是合理的，但这里有一个替代方案：

import collections
with open('grid_file.txt', 'r') as f:
    columns = next(f).split()
    matrix = collections.defaultdict(dict)
    for line in f:
        items = line.split()
        row, vals = items[0], items[1:]
        for col, val in zip(columns, vals):
            matrix[col][row] = int(val)
print(matrix)

产生

defaultdict(<type 'dict'>, {'A': {'A': 0, 'C': 6, 'B': 3}, 'C': {'A': 2, 'C': 0, 'B': 5}, 'B': {'A': 1, 'C': 7, 'B': 0}})

一些提示：

Use
```
with open(...) as f
    ...
```
instead of
```
f = open(...)
f.close()
```
because the file handle is closed for you when Python leaves the with-block. By using with you'll never forget to close a filehandle, and even if an exception occurs, the filehandle will still be closed upon leaving the with-block.
Generally, it is better to avoid f.readlines() if you can. This slurps the entire file into a list. That can be onerous on memory, especially if the file is huge. Usually
```
with open(...) as f:
    for line in f:
```
can be used instead.
If you make matrix a collections.default(dict) then matrix[field] will be a dict by default. So you can skip the initialization:
```
for key in keys:
    matrix[key] = {}
```
A defaultdict is a subclass of dict, so you can use it very much as you would a dict. If you don't like the way it prints or would like to stop matrix from automagically assigning an empty dict to matrix[key] for any key, you can convert the defaultdict back to a regular dict with:
```
matrix = dict(matrix)
```
Avoid using numerical indices in for-loops if you can.
```
for i, c in enumerate(chars[1:]):
```
Although this is de rigueur for most C-like languages, Python has a better way: looping over the items themselves:
```
for col, val in zip(columns, vals):
```
This makes the code more readable, because it assigns a variable name to the object you are actually interested in, not just an index which you then have to compose into things like keys[i-1]. It also helps you avoid "off-by-one" errors which can occur when you have to adjust the index by one, as is done in keys[i-1].

另一种可能是不使用嵌套字典，而是将2元组（列，行）作为键：

with open('grid_file.txt', 'r') as f:
    columns = next(f).split()
    matrix = {}
    for line in f:
        items = line.split()
        row, vals = items[0], items[1:]
        for col, val in zip(columns, vals):
            matrix[col, row] = int(val)
print(matrix)

产量

{('B', 'C'): 7, ('A', 'A'): 0, ('B', 'B'): 0, ('B', 'A'): 1, ('C', 'A'): 2, ('C', 'B'): 5, ('C', 'C'): 0, ('A', 'B'): 3, ('A', 'C'): 6}

然后你可以像这样访问矩阵中的 (列, 行)：

print(matrix['A','C'])
# 6

顺便提一下，如果你安装了 pandas：

import pandas as pd
import io

text = '''\
A  B  C
A 0  1  2
B 3  0  5
C 6  7  0'''

df = pd.read_table(io.BytesIO(text), sep='\s+')
print(df.to_dict())

产出

{'A': {'A': 0, 'B': 3, 'C': 6},
 'B': {'A': 1, 'B': 0, 'C': 7},
 'C': {'A': 2, 'B': 5, 'C': 0}}