CsvMapper and @JsonFilter annotation

[mablaev]

CsvMapper does not generate valid output, when I trying to use it against bean with @jsonfilter annotation.
There are 2 problems here:

1. I use csv schema with header generation. Filter do not affect on header.
2. JsonFilter damages csv content by removing coma column separator.

public class AppTest {

@Test
public void testCsvMapperWithJsonFilter() throws JsonProcessingException {
    CsvMapper mapper = new CsvMapper();
    CsvSchema schema = mapper.schemaFor(Company.class).withLineSeparator("\n").withHeader();

    SimpleFilterProvider filterProvider = new SimpleFilterProvider()
            .addFilter(COMPANY_FILTER, FilterExceptFilter.filterOutAllExcept("name", "ticker"));

    String actual = mapper.writer(filterProvider).withSchema(schema).writeValueAsString(createCompanies());
    System.out.println(actual);

    StringBuilder expected = new StringBuilder("name,ticker")
            .append('\n')
            .append("name1,ticker1")
            .append('\n')
            .append("name2,ticker2")
            .append('\n')
            .append("name3,ticker3");

    Assert.assertEquals(expected, actual);
}

private static List<Company> createCompanies() {
    return Arrays.asList(
            new Company(1, "name1", "ticker1")
            , new Company(2, "name2", "ticker2")
            , new Company(3, "name3", "ticker3"));
}

@JsonFilter(COMPANY_FILTER)
public static class Company {
    static final String COMPANY_FILTER = "COMPANY_FILTER";

    private int id;
    private String name;
    private String ticker;

    private Company(int id, String name, String ticker) {
        this.id = id;
        this.name = name;
        this.ticker = ticker;
    }

    public int getId() {
        return id;
    }

    public void setId(int id) {
        this.id = id;
    }

    public String getName() {
        return name;
    }

    public void setName(String name) {
        this.name = name;
    }

    public String getTicker() {
        return ticker;
    }

    public void setTicker(String ticker) {
        this.ticker = ticker;
    }
}

}

[cowtowncoder]

I will try to solve this as far as possible.

One part that may be difficult is that of doing anything to header line: I realize that ideally header columns would be removed. This may be difficult to handle mostly because data-binder has no knowledge of special handling of data formats.

But even if headers could not be filtered, it should be possible to add placeholders (empty values) for missing data.

[cowtowncoder]

Ok: I implemented at least partial support for both @JsonView and @JsonFilter-based filtering. In both cases, values of properties indicated to be filtered out are omitted and empty value is included instead. Header line is left as-is, for this version as I could not yet figure out how to exclude it.

What I am thinking is that this baseline support will allow some level of filtering; and the additional (possibly optional) feature of excluding column name and value completely from output can and should be filed as a separate feature request. This is mostly so that release notes will include exact information of what is included and works.

[mablaev]

thank you!