含特殊字符的Cortex型号正则匹配失效，请求优化正则表达式

Hey there, I see the problem—your current regex fails to capture the full Cortex-AXX pattern when trademark symbols (like ® or ™) or other special characters sit between Cortex and the model number. Let's tweak the regex to handle those pesky interfering characters properly.

Improved Regex Pattern #

Here's a revised regex that skips over any non-alphanumeric, non-hyphen characters between the parts of the model name:

(?i)(Cortex)\b[^\w-]*-?[^\w-]*(\bA\b)[^\w-]*-?[^\w-]*(\d{1,2})

If you want something a bit more concise (and still effective):

(?i)Cortex\s*[^\w]*-?[^\w]*A[^\w]*-?[^\w]*(\d{1,2})

Breakdown of the Regex #

Let's walk through each part so you understand how it works:

• (?i): Turns on case-insensitive matching, so it catches cortex, Cortex, CORTEX—no matter the capitalization.
• (Cortex)\b: Captures the core name, with a word boundary (\b) to avoid accidental matches like CortexX.
• [^\w-]*: Matches any number of characters that aren't letters, numbers, or hyphens—this covers those ®, ™, random spaces, or other special symbols messing up your match.
• -?: Matches an optional hyphen, since some models are written as Cortex A53 and others as Cortex-A53.
• (\bA\b): Captures the A prefix with word boundaries to ensure we're not matching part of another word.
• (\d{1,2}): Grabs the 1-2 digit model number (like 53, 54, etc.).

Testing It Out #

Let's test this against your example inputs:

1. Clean input: Helio P10 processor ARM cortex-A53 1.8GHz
   • Match result: cortex-A53 (with capture groups for cortex, A, and 53)
2. Input with special chars: Helio P10 processor ARM® cortex®-A53™1.8GHz
   • Match result: cortex®-A53 (if you want to strip the special symbols afterward, you can use a simple string replace, or adjust the regex to only capture the core parts and reconstruct the clean format)

If you want to directly get a clean Cortex-AXX string without any special symbols, use this regex and combine the capture groups:

(?i)(Cortex)\b.*?(\bA\b).*?(\d{1,2})

Then just concatenate the groups with a hyphen: $1-$2$3 will give you Cortex-A53 even from the messy input.

Quick Tip #

If you run into other special characters (like underscores or periods), you can adjust the [^\w-] part to [^A-Za-z0-9-] to cover all bases. For broader Unicode support (if dealing with non-Latin special chars), use \P{L} instead to match any non-letter characters—it works better in regex engines that support Unicode properties.

内容的提问来源于stack exchange，提问作者Bhagwati Malav