EDGAR文件中ITEM标记间文本提取的正则贪婪/非贪婪匹配问题

Got it, let's break down why your current regex is failing and fix it for reliable EDGAR filing text extraction.

The Core Issue #

Your regex item[^a-zA-Z\n]*\d\s*\.\s*management\'s discussion and analysis.*?item[^a-zA-Z\n]*\d\s* has two critical flaws:

• Non-greedy .*? stops at the first mention of any item X in the body text (like a reference to "Item 1" in a sentence), not the actual next top-level ITEM heading.
• Greedy .* overshoots and grabs everything up to the last ITEM tag in the document, including unrelated sections.

EDGAR's top-level ITEM headings have consistent, identifiable patterns—we can leverage that to avoid false matches.

The Solution: Target Only Top-Level ITEM Headings #

Top-level ITEMs in EDGAR filings almost always:

• Start at the beginning of a line
• Use all caps (or mixed case) for ITEM
• Follow the format ITEM [Number][Optional Letter]. [Section Title]
• Often have separator lines (like --- or ===) below them

Here's a regex that targets these patterns, so it only stops at the next valid top-level ITEM:

Python Example Code

import re

# Load your EDGAR text file
with open("edgar_filing.txt", "r", encoding="utf-8") as file:
    edgar_text = file.read()

# Regex to extract ITEM 1 (MDA) up to the next top-level ITEM
mda_regex = r"""(?ix)
    (^ITEM\s+(\d+)(?:[A-Z])?\s*\.\s*Management's Discussion and Analysis.*?)  # Match ITEM 1 heading + separator
    (.*?)                                                                     # Non-greedy match content
    (?=^ITEM\s+(?!\2)\d+(?:[A-Z])?\s*\.)                                      # Stop at next top-level ITEM with different number
"""

# Run the match with flags: DOTALL (match newlines), MULTILINE (^ matches line starts)
matches = re.findall(mda_regex, edgar_text, re.DOTALL | re.MULTILINE)

if matches:
    # Extract the actual MDA content (third group in the match)
    mda_content = matches[0][2].strip()
    print(mda_content)
else:
    print("No valid MDA section found.")

Key Improvements Explained #

1. Line-Start Matching: Using ^ with re.MULTILINE ensures we only match ITEMs that are actual headings, not random mentions in the body.
2. Reverse Reference Check: (?!\2) ensures we don't stop at a sub-item (like ITEM 1.1) or a reference to the same ITEM number in the text—we only stop at the next top-level ITEM with a different number.
3. Flexible Format Support: The regex handles optional letters (like ITEM 1A) and separator lines below the heading.
4. Readable Regex: The (?x) flag lets us add whitespace/comments to the regex for clarity.

For Exact ITEM Ranges #

If you need to extract between specific ITEMs (e.g., ITEM 1 to ITEM 2), simplify the regex to target those exact numbers:

specific_range_regex = r'(?i)(^ITEM\s+1\s*\.\s*Management\'s Discussion and Analysis.*?)(.*?)(?=^ITEM\s+2\s*\.)'

This will lock onto the exact section you need without any false stops.

内容的提问来源于stack exchange，提问作者Rohit Haritash