python - 对单词进行计数和排名时，Python 中的大型 .txt 文件出现 MemoryError

Question

我正在尝试从包含芬兰语文本的 500mb 文本文件创建排名单词列表 csv 文件。该脚本可以处理小文件，但不适用于 500mb 的文件。

我是 Python 的初学者，所以如果它相当马虎，请原谅我。环顾四周，我想我可能必须逐行处理文件。

with open(...) as f:
    for line in f:
    # Do something with 'line' 

我将不胜感激任何指点，干杯!我的代码如下:

#load text
filename = 'finnish_text.txt'
file = open(filename, 'r')
text = file.read()
file.close()

#lowercase and split words by white space
lowercase = text.lower()
words = lowercase.split()

# remove punctuation from each word
import string
table = str.maketrans('', '', string.punctuation)
stripped = [w.translate(table) for w in words]

# ranked word count specify return amount here
from collections import Counter
Counter = Counter(stripped)
most_occur = Counter.most_common(100)

# export csv file
import csv
with open('word_rank.csv', 'w', newline='') as csvfile:
    writer = csv.writer(csvfile, delimiter=',')
    for x in most_occur:
        writer.writerow(x)

编辑:我最终使用了@Bharel(多么传奇)在他的评论中给出的第二个解决方案。由于编码问题，我不得不更改几行。

with open(filename, 'r', encoding='Latin-1', errors='replace') as file:

with open('word_rank.csv', 'w', newline='', errors='replace') as csvfile:

Answer 1

将所有内容切换到生成器，它应该可以工作:

#load text
filename = 'finnish_text.txt'
# Auto-close when done
with open(filename, 'r') as file:

    #lowercase and split words by white space
    word_iterables =(text.lower().split() for line in file)

    # remove punctuation from each word
    import string
    table = str.maketrans('', '', string.punctuation)

    stripped = (w.translate(table) for it in word_iterables for w in it)

    # ranked word count specify return amount here
    from collections import Counter
    counter = Counter(stripped)

most_occur = counter.most_common(100)

# export csv file
import csv
with open('word_rank.csv', 'w', newline='') as csvfile:
    writer = csv.writer(csvfile, delimiter=',')
    for x in most_occur:
        writer.writerow(x)

通过使用生成器(括号而不是方括号)，所有单词都会被延迟处理，而不是一次全部加载到内存中。

<小时/>

如果您想要最有效的方法，我已经写了一个作为 self 挑战:

import itertools
import operator

#load text
filename = 'finnish_text.txt'
# Auto-close when done
with open(filename, 'r') as file:

    # Lowercase the lines
    lower_lines = map(str.lower, file)

    # Split the words in each line - will return [[word, word], [word, word]]
    word_iterables = map(str.split, lower_lines)

    # Combine the iterables:
    # i.e. [[word, word], [word, word]] -> [word, word, word, word]
    words = itertools.chain.from_iterable(word_iterables)

    import string
    table = str.maketrans('', '', string.punctuation)

    # remove punctuation from each word
    stripped = map(operator.methodcaller("translate", table), words)

    # ranked word count specify return amount here
    from collections import Counter
    counter = Counter(stripped)

most_occur = counter.most_common(100)

# export csv file
import csv
with open('word_rank.csv', 'w', newline='') as csvfile:
    writer = csv.writer(csvfile, delimiter=',')
    for x in most_occur:
        writer.writerow(x)

它充分利用了用 C 编写的生成器(map 和 itertools)。