如何将编号列表切片为子列表

Question

我打开了一个文件，并使用正则表达式't' readlines()和split()来删除TAB，它导致了以下列表：

["1", "cats", "--,"]
["2", "chase", "--,"]
["3", "dogs", "--,"]
["1", "the", "--,"]
["2", "car", "--,"]
["3", "is", "--,"]
["4", "gray", "--,"]

现在我想通过循环索引 [0] 上的整数作为句子边界来提取并切成子列表，例如"猫追狗"和"汽车是灰色的"。例如，1 - 3 子列表"猫追狗"，然后继续计数 1 - 4 子列表"汽车是灰色的"，依此类推，以便我得到子列表["the", "car", "is", "gray" ]。我该怎么做？

我已经试过了，但出现错误：

无法连接 int + str

将 for 循环中的"i"检测为字符串元素而不是整数：

with open(buffer, 'r') as f:
    words = []
    for line in f:
        items = line.split('t')[:1]
        for i in items:
            while i>1:
                i = i+1
                print i

Answer 1

像这样：

from itertools import groupby
with open('yourfile') as fin:
    # split lines
    lines = (line.split() for line in fin)
    # group by consecutive ints
    grouped = groupby(enumerate(lines), lambda (idx, el): idx - int(el[0]))
    # build sentences from words in groups
    sentences = [' '.join(el[1][1] for el in g) for k, g in grouped]
    # ['cats chase dogs', 'the car is gray']

注意：这基于您的示例数据：

example = [
    ["1", "cats", "--,"],
    ["2", "chase", "--,"],
    ["3", "dogs", "--,"],
    ["1", "the", "--,"],
    ["2", "car", "--,"],
    ["3", "is", "--,"],
    ["4", "gray", "--,"]
]

Answer 2

选择合适的数据结构使工作更容易：

container = [["1", "cats", "--,"],
             ["2", "chase", "--,"],
             ["3", "dogs", "--,"],
             ["1", "the", "--,"],
             ["2", "car", "--,"],
             ["3", "is", "--,"],
             ["4", "gray", "--,"]]

将列表嵌套在容器列表中，然后使用字典存储输出列表：

from collections import defaultdict
out = defaultdict(list)              # Initialize dictionary for output
key = 0                              # Initialize key  
for idx, word, _ in container:       # Unpack sublists
    if int(idx) == 1:                # Check if we are at start of new sentence
        key += 1                     # Increment key for new sentence
    out[key].append(word)            # Add word to list

给：

{
    1: ['cats', 'chase', 'dogs'], 
    2: ['the', 'car', 'is', 'gray']
}