Python 解析大文本文档，仅保留；帐号；，和特定关键字（“市场价值”）

我有一个大的文本文档（~20000行），其主体如下所示：

Invoice Account / Name: 
0234523454 / XYZCORPORATIONS
Charge Group
Portfolio Fee
Date
Our / Your Ref
Security / Category
Charge Item
No of Units
Market Value
Charge Amt Invoice Amt
30-Sep-2019
Debt Instruments
PORTFOLIO FEE
CS
USD 
USD 219.12 USD 219.12
14,136,666.31
 Invoice Account / Name: 
021346676343/ abcdefgcopr
M0919-031  / Page 3 of 35
Charge Group
Portfolio Fee
Date
Our / Your Re
Security / Category
Charge Item
No of Units
Market Value
Charge Amt Invoice Amt
30-Sep-2019
Equity Instruments
USD 788,640.00 USD 12.22
USD 12.22
PORTFOLIO FEE-
EC_CS
 Invoice Account / Name: 
123498761233/ somethingelsecorporation
Charge Group
Portfolio Fee
Date
Our / Your Ref

Invoice Account / Name: 
0234523454 / XYZCORPORATIONS

像这样的块重复了数千次。 正在尝试输出：

Invoice Account / Name: 
    0234523454 / XYZCORPORATIONS
Market Value
Invoice Account / Name: 
    021346676343/ abcdefgcopr
Market Value
Invoice Account / Name: 
    123498761233/ somethingelsecorporation
Market Value

由于我以前从未尝试过这样的事情，我有两个问题：
1.如何识别和保留这样的句子：

Invoice Account / Name: 
0234523454 / XYZCORPORATIONS
Charge Group
Portfolio Fee
Date
Our / Your Ref
Security / Category
Charge Item
No of Units
Market Value
Charge Amt Invoice Amt
30-Sep-2019
Debt Instruments
PORTFOLIO FEE
CS
USD 
USD 219.12 USD 219.12
14,136,666.31
 Invoice Account / Name: 
021346676343/ abcdefgcopr
M0919-031  / Page 3 of 35
Charge Group
Portfolio Fee
Date
Our / Your Re
Security / Category
Charge Item
No of Units
Market Value
Charge Amt Invoice Amt
30-Sep-2019
Equity Instruments
USD 788,640.00 USD 12.22
USD 12.22
PORTFOLIO FEE-
EC_CS
 Invoice Account / Name: 
123498761233/ somethingelsecorporation
Charge Group
Portfolio Fee
Date
Our / Your Ref

Invoice Account / Name: 
0234523454 / XYZCORPORATIONS

哪个没有固定长度

如何只保留关键字“市场价值”而不是这个

---

为此使用nltk明智吗？或者可以用正则表达式和字符串处理来处理它

您可以使用字符串处理来搜索并找到您正在寻找的内容

result = []

with open('num.txt', 'r') as file:
    data = list(file.readlines())

    for indx, row in enumerate(data): 
        if 'Invoice Account' in row:

            accountnumber = data[indx+1].split('/')[0].strip() # Get account number from next line
            companyname = data[indx+1].split('/')[1].strip() # Get company name from next line

            # Store all results in a dictionary, you could print, store in other ways as well.

            info = {'Account Number': accountnumber,
                    'Company Name': companyname,
                    'Market Value': '',
                }

            # Append the dictionary to a list called result
            result.append(info)

然后，您可以直接从每个字典访问数据，其中只包含单个公司的值

for data in result:
    print(f"""Account Name: {data['Company Name']}
Account Number: {data['Account Number']}
Market Value: {data['Market Value']}
""")

输出：

Account Name: XYZCORPORATIONS
Account Number: 0234523454
Market Value: 

Account Name: abcdefgcopr
Account Number: 021346676343
Market Value: 

Account Name: somethingelsecorporation
Account Number: 123498761233
Market Value: 

---

您只能使用

regex

实现它：

import re

with open('file.txt', 'r') as f:
    matches = re.findall('Invoice Account \/ Name:\n.*', f.read())

with open('result.txt', 'w') as f:
    [f.write(f'{m}\nMarket Value\n') for m in matches]

输出文件：

Invoice Account / Name:
0234523454 / XYZCORPORATIONS
Market Value
Invoice Account / Name:
021346676343/ abcdefgcopr
Market Value
Invoice Account / Name:
123498761233/ somethingelsecorporation
Market Value

---

您是要存储实际市场价值，还是只存储文本

市场价值

？只存储文本市场价值。我不需要任何其他号码，除了帐号。到底是什么问题？你试过什么，做过什么研究吗？在更大的数据集上，这比我的答案快得多+1@Alderven，感谢您迄今为止的解决方案。但这假设每个“发票账户/名称”之间只有一个“市场价值”关键字。然而，在txt文件中，有时会多次出现“市场价值”在两个账号之间<代码>发票账户/名称：0234523454/XYZ公司市值市值发票账户/名称：021346676343/abcdefgcopr市值发票账户/名称：123498761233/Somethingelscorporation市值市值市值为没有早些时候解释这一错误表示歉意，新手错误。