arrays - 在 pyspark 中展平嵌套的 Json，其中每个数组类型都是一个单独的数据框

Question

以这个模式（json数据框）为例：

 root
     |-- abstract: string (nullable = true)
     |-- adx_keywords: string (nullable = true)
     |-- asset_id: long (nullable = true)
     |-- byline: string (nullable = true)
     |-- column: string (nullable = true)
     |-- des_facet: array (nullable = true)
     |    |-- element: string (containsNull = true)
     |-- eta_id: long (nullable = true)
     |-- geo_facet: array (nullable = true)
     |    |-- element: string (containsNull = true)
     |-- id: long (nullable = true)
     |-- media: array (nullable = true)
     |    |-- element: struct (containsNull = true)
     |    |    |-- approved_for_syndication: long (nullable = true)
     |    |    |-- caption: string (nullable = true)
     |    |    |-- copyright: string (nullable = true)
     |    |    |-- **media-metadata**: array (nullable = true)
     |    |    |    |-- element: struct (containsNull = true)
     |    |    |    |    |-- format: string (nullable = true)
     |    |    |    |    |-- height: long (nullable = true)
     |    |    |    |    |-- url: string (nullable = true)
     |    |    |    |    |-- width: long (nullable = true)
     |    |    |-- subtype: string (nullable = true)
     |    |    |-- type: string (nullable = true)
     |-- nytdsection: string (nullable = true)
     |-- org_facet: array (nullable = true)
     |    |-- element: string (containsNull = true)
     |-- per_facet: array (nullable = true)
     |    |-- element: string (containsNull = true)
     |-- published_date: string (nullable = true)
     |-- section: string (nullable = true)

在这里，我想将每个数组类型作为单独的数据框。我能够分离第一级，如 des_facet、geo_facet 和媒体，并将其放入单独的数据帧中，但无法对第二级执行相同操作，例如： 媒体中的媒体元数据（数组类型）。将嵌套数组提取为单个数据框的解决方案将是有用的

提前致谢