Stripping unicode null character from the TextNode

[pwspot]

I am looking for the proper way to strip unicode null (\u0000) characters from TextNode objects, within the ObjectNode (extending JsonNode).

Firstly, a JSON String containing \\u0000 (double backslash is intentional here) is converted into a JsonNode. The result object contains a LinkedHashMap with an element with a key of type String and a value of type TextNode. The TextNode unescapes the String and it actually contains a single unicode null character.

Secondly, the resulting object can become a property of another object or an element of a collection. Then this "another object" or a collection is serialized using valueToTree. Here, for simplicity, I am skipping this "wrapping" part and it's just calling valueToTree over the JsonNode:

String jsonWithUnicodeNull = "{ \"foo\": \"\\u0000\" }";
ObjectMapper mapper = new ObjectMapper();

// 1
JsonNode node = mapper.readTree(jsonWithUnicodeNull);

// 2
ObjectNode out = mapper.valueToTree(node);

What I am trying to achieve is to strip the unicode null characters when calling the valueToTree over the 2nd object.
I could not find any way to apply a JsonSerializer to work with the TextNode objects.
I was able to do that by overriding the mapper's JsonNodeFactory:

mapper.setNodeFactory(new JsonNodeFactory(false) {
  @Override
  public TextNode textNode(String text) {
    return new TextNode(org.apache.commons.lang3.StringUtils.remove(text, '\u0000'));
  }
});

What I don't like with this approach is the assumption on what the original node factory object is and its defaults.

For the context, these are another settings that I use with the mapper object:

mapper.configure(SerializationFeature.EAGER_SERIALIZER_FETCH, true); // is default
mapper.configure(SerializationFeature.FAIL_ON_EMPTY_BEANS, false);
mapper.setSerializationInclusion(Include.NON_NULL);
mapper.configure(DeserializationFeature.FAIL_ON_UNKNOWN_PROPERTIES, false);

I am not stripping the character in the first step (readTree) as this actually can come from multiple sources and filtering it in a single place seems more reasonable.

Version information

Which Jackson version(s) was this for? 2.14.2
JDK - Temurin-11.0.15+10

[cowtowncoder]

My suggestion would be read content as JsonNode and then post-process it by traversing, to change TextNodes where necessary.

[pwspot]

Could you guide me to any example of post-processing?

[cowtowncoder]

Just traverse over values recursively (need to handle ArrayNode and ObjectNode specifically), look for TextNodes, replace if change needed.