如果文件存在，则使用正则表达式提取 json 日志中的字段答案

【问题标题】：Extract the fields in json log if the filed exists using regex如果文件存在，则使用正则表达式提取 json 日志中的字段
【发布时间】：2019-11-19 05:55:45
【问题描述】：

我正在尝试提取以下示例数据中的字段

{"Value":{"Data":{"Items":[{"FieldType":"ABC","Value":"*****"},{"FieldType":"ACB","Value":"*****"},{"FieldType":"ABCD","Value":"*****"},,{"FieldType":"XYZ","Value":"*****"}],"EncryptedDocKey":"*****","Domain":"*****","Partner":"***","Carrier":"****"},"RequestTrackerId":"****","Message":"OK"},"Formatters":[],"ContentTypes":[],"DeclaredType":null,"StatusCode":null}

{"Value":{"Data":{"Items":[{"FieldType":"ABC","Value":"*****"},{"FieldType":"XYZ","Value":"*****"},{"FieldType":"ACD","Value":"*****"}],,"EncryptedDocKey":"*****","Domain":"*****","Partner":"***","Carrier":"****"},"RequestTrackerId":"****","Message":"OK"},"Formatters":[],"ContentTypes":[],"DeclaredType":null,"StatusCode":null}

{"Value":{"Data":{"Items":[{"FieldType":"XYZ","Value":"*****"},{"FieldType":"ACD","Value":"*****"}],,"EncryptedDocKey":"*****","Domain":"*****","Partner":"***","Carrier":"****"},"RequestTrackerId":"****","Message":"OK"},"Formatters":[],"ContentTypes":[],"DeclaredType":null,"StatusCode":null}

{"Value":{"Data":{"Items":[{"FieldType":"ABCD","Value":"*****"}],,"EncryptedDocKey":"*****","Domain":"*****","Partner":"***","Carrier":"****"},"RequestTrackerId":"****","Message":"OK"},"Formatters":[],"ContentTypes":[],"DeclaredType":null,"StatusCode":null}

我需要将FieldType (ABC,XYZ,ABCD,ACB,ACB ...)旁边的文字提取到V1、V2、V3、V4、V5等...

如果有三个字段，我有一个正则表达式可以获取，但如果字段多于或少于 3 个，则正则表达式不会映射。

我试过这个正则表达式

.?"FieldType":"(?<V1>\w+)".+?"FieldType":"(?<V2>\w+).+?"FieldType":"(?<V3>\w+)

我还想提取fieldtype和,EncryptedDocKey,Domain,Partner,Carrier,RequestTrackerId,Message,Formatters,ContentTypes,DeclaredType,StatusCode中的所有字段

【问题讨论】：

为什么不使用 JSON.NET？
我强烈建议您不要自己解析 JSON，当然也不要使用正则表达式。这是我写的一篇关于这个主题的文章：blog.newrelic.com/engineering/7-things-never-code

标签： regex regex-lookarounds regex-negation regex-group regex-greedy

【解决方案1】：

这里没有必要应用正则表达式，但如果必须，可以使用类似于：

"(FieldType)":"([^"]+)"|"(message)":"([^"]+)"

可能会提取那些所需的键值。

DEMO

【讨论】：