正则表达式提取设备型号问题：修复User-Agent匹配异常

The issue with your current regex .*?\(.*;(?<model>.*?)(; wv)?\) is that the lazy .*? matcher prioritizes the shortest possible match when it hits the optional (; wv)? group. For your third User-Agent example, this causes it to capture "wv" instead of the intended model string. Let's adjust the regex to explicitly handle both scenarios (with and without ; wv at the end of the parenthetical content):

Corrected Regex #

.*?\(.*?(?:;(?<model>[^;]+); wv|;(?<model>[^;]+))\)

Regex Breakdown #

This pattern aligns perfectly with your extraction rules:

• .*?\(: Non-greedily matches up to the opening parenthesis of the first (...) section.
• .*?: Non-greedily matches content inside the parentheses until we hit one of two target patterns.
• (?:...): A non-capturing group that handles both cases:
  1. ;(?<model>[^;]+); wv: For entries ending with ; wv, this captures all characters (no internal semicolons) between the penultimate semicolon and the ; wv suffix.
  2. ;(?<model>[^;]+): For entries without ; wv, this captures all characters after the final semicolon.

Test Results Against Your Examples #

1. First User-Agent: Mozilla/2.0 (Linux; Android 2.0; KK; SAMSUNG SM-G900H Build/LRX21T) AppleWebKit
   • Captures: SAMSUNG SM-G900H Build/LRX21T
2. Second User-Agent: Mozilla/5.0 (Linux; Android 5.0; SAMSUNG SM-G900H Build/LRX21T) AppleWebKit
   • Captures: SAMSUNG SM-G900H Build/LRX21T
3. Third User-Agent: Mozilla/5.0 (Linux; Android 5.1; XT1032 Build/LPBS23.13-56-2; wv) AppleWebKit
   • Captures: XT1032 Build/LPBS23.13-56-2

内容的提问来源于stack exchange，提问作者Giorgi Nakeuri