单个正则表达式检索多个“内部”文本（或子字符串）

Question

I'm looking for an elegant way to handle regex of content where there are multiple "headers" and then multiple repeated "inner keys". I'm wondering if this is possible or if what I'm doing (ie pull out each group and then process each inner data set) is the best method.

I've been able to do the "brute force" (aka two regex) method where I do a regex pass to get the header content and block of key data; then do a second pass on each of the key data blocks with a second regex to pull out those details; but I'm wondering if there is a way to do a single regex that would have one match for each key, but also include the header data?

var headerRegex = new Regex(@"HEADER.+?name = (?<name>[\w\s]+?)detail(?<keyData>.+?)HEADER", RegexOptions.Singleline);
var keyRegex = new Regex(@"KEY.+?name = (?<name>[\w\s]+?)type", RegexOptions.Singleline);
foreach (Match headerMatch in headerRegex.Matches(input))
{
   foreach (Match keyMatch in keyRegex(headerMatch.Groups["keyData"].Value))
   {
     // here I have header and key data
     // goal: to have a single foreach with a regex that has head and key data
   }
}

The content I'm working with has the form (note: it is not JSON formatted, more "json-esque")

HEADER
{
  name = name content
  detail = detail content
  sample = sample content
  KEY
  {
    name = name content
    type = type content
    value = value content
  }
  KEY
  {
    name = name content
  }
  KEY
  {
    name = name content
  }
  additional = additional content
  more = more content
}
HEADER
{
  name = name content
  detail = detail content
  sample = sample content
  KEY
  {
    name = name content
    type = type content
    value = value content
  }
  KEY
  {
    name = name content
  }
  KEY
  {
    name = name content
  }
  additional = additional content
  more = more content
}

Answer 1

Based on your current regex, looks like you intend to match only the values for name keys.

Option 1: Simplest option would be https://regex101.com/r/jRRwsv/1 .

Pattern: (?:HEADER|KEY).+?name = (?<name>[\w\s]+?)\n

Explanation: String should either start with HEADER or KEY , followed by name . This will not differentiate between the name values in HEADER and KEY section.

---

Option 2: To capture HEADER and KEY names separately, one option is to split the expand the regex as (?:HEADER.+?name = (?<hname>[\w\s]+?)\n)|(?:KEY.+?name = (?<kname>[\w\s]+?)\n) .

Refer https://regex101.com/r/jRRwsv/2

---