Add methods for appending columns of a `CsvSchema` into another

[sharpau]

Currently if I am building a CsvSchema with columns from a POJO and some other custom columns, I have to first build a CsvSchema from the POJO, and then provide the columns from that schema to a second CsvSchema.Builder. It's not a huge deal, but it seems like it'd be a nice improvement.

Current situation:

        CsvMapper mapper = new CsvMapper();
        CsvSchema pojoSchema = mapper.schemaFor(Pojo.class);

        CsvSchema fullSchema = CsvSchema.builder()
                .addColumns(pojoSchema)
                .addColumns(otherProperties, CsvSchema.ColumnType.STRING)
                .build().withHeader().withNullValue("");

Desired situation:

        CsvMapper mapper = new CsvMapper();

        CsvSchema fullSchema = CsvSchema.builder()
                .addColumnsFromPojo(Pojo.class)
                .addColumns(otherProperties, CsvSchema.ColumnType.STRING)
                .build().withHeader().withNullValue("");

[cowtowncoder]

While I understand why it would be convenient, I am not sure it is feasible because actual schema generation requires knowledge that only ObjectMapper has (and/or perhaps ObjectReader or ObjectWriter) -- basically information equivalent to knowing how to serialize a POJO.

It might be possible to add ObjectMapper argument.
Or add specific method in CsvMapper to make this more convenient... like, give CsvSchema.Builder initialized with columns for given POJO? That might be the way to go.

[cowtowncoder]

Added CsvSchema.withColumnsFrom(CsvSchema) and CsvSchema.Builder.addColumnsFrom(CsvSchema) methods which should help here.
It is unfortunately not possible to pass just type (Class, JavaType) since these objects do not have proper execution context (ObjectMapper or ObjectWriter) for introspection; but I think this should be helpful enough considering that mapper/writer is needed for initial introspection anyway.