How does Spark in Java filter the values in the list in dataset?

I have two class, one is NewsArticle: String id, String title, List<ContentItem> contents, the other is ContentItem: String content, String subtype, String url.

I want to filter out the content whose subtype value is equal to "paragraph", and spliced into one long string. (don't need url)

here is the NewsArticle Dataset like:

 1, "TiTle", [{htt..., paragraph, rem...},{htt..., paragraph, rem...},{htt..., paragraph, rem...}]

which is id, title, List<ContentItem>

I took out the contents column, and each single row is one article, it like this:

[{http..., others, con...},{http..., paragraph, rem...},{http..., paragraph, rem...}]

which is url, subtype, content

and now I want to make each article(row) look like:

1, "Title", "this is content which subtype equals paragraph"

can anyone help me with java?

This would work:

df
     .withColumn("newContent", functions.explode(functions.col("items")))
     .filter("newContent.subtype=='paragraph'")
     .selectExpr("id", "title", "newContent.content as content")
     .show(false);

Input:

+---+--------------------------------------------------------------------------------------------------------+-----+
|id |items                                                                                                   |title|
+---+--------------------------------------------------------------------------------------------------------+-----+
|id |[[Content1, subtype1, someurl], [ContentOfParagraph, paragraph, someurl], [Content2, subtype2, someurl]]|Title|
+---+--------------------------------------------------------------------------------------------------------+-----+

Output:

+---+-----+------------------+
|id |title|content           |
+---+-----+------------------+
|id |Title|ContentOfParagraph|
+---+-----+------------------+