如何防止CsvMapper在输出CSV中包含无法识别的字段？

Question

When using CsvMapper with @JsonPropertyOrder , all fields that are not explicitly named are put at the end of the CSV record. This means that typos result in different CSV ordering.

Example:

public class Example {
  public static void main(String[] args) throws JsonProcessingException {
    final CsvMapper csvMapper = new CsvMapper();
    System.out.println(csvMapper.writerWithSchemaFor(MyClass.class).writeValueAsString(new MyClass("test", true, 2)));
  }

  @Data
  @AllArgsConstructor
  @NoArgsConstructor
  @JsonPropertyOrder({"name", "code"})
  public static class MyClass {

    @JsonProperty("name")
    private String name;

    @JsonProperty("status")
    private boolean status;

    @JsonProperty("code")
    private Integer code;
  }
}

Results in: test,2,true . Since status was not explicitly named in @JsonPropertyOrder it was put in the end of the record. I would prefer if CsvMapper threw an exception or if it was not serialized at all.

Another example would be if I changed @JsonPropertyOrder above to @JsonPropertyOrder({"name", "stats", "code"}) . Notice the typo in stats . This would result in the same CSV as above.

How can I configure CsvMapper to either fail on unrecognized properties or not serialize properties not named in @JsonPropertyOrder ?

Answer 1

IMHO the only purpose of @JsonPropertyOrder is to give you a chance of defining an order of your properties. Regarding the unspecified fields you could only use the alphabetic attribute to control their ordering. Consequently the annotation ignores typos and just cares about the known field names you've specified in its value attribute.

If you need a more granular control of active fields for (de-)serialization you could use @JsonIgnoreProperties on your class or @JsonIgnore on your fields.

May be this analogy is helpful: Consider an SQL statement – you SELECT the rows in a table and ORDER the result set afterwards. Here the ORDER command has also no influence on the result set's length or width.