TCK schema is not complete when generated with AvroToRecord tool

I have an avro payload (see attached file customers_ordres.avro) and i try to convert it to TCK record. I use for that the org.talend.components.common.stream.input.avro.AvroToRecord tool.
The input avro payload contains a "customer" array with records. The schema of these records is the following:

{
	"type": "record",
	"name": "customer",
	"fields": [{
		"name": "custid",
		"type": ["null",
		"string"],
		"default": null
	},
	{
		"name": "name",
		"type": ["null",
		"string"],
		"default": null
	},
	{
		"name": "address",
		"type": ["null",
		{
			"type": "record",
			"name": "address",
			"fields": [{
				"name": "street",
				"type": ["null",
				"string"],
				"default": null
			},
			{
				"name": "city",
				"type": ["null",
				"string"],
				"default": null
			},
			{
				"name": "zipcode",
				"type": ["null",
				"string"],
				"default": null
			}]
		}],
		"default": null
	},
	{
		"name": "rating",
		"type": ["null",
		"int"],
		"default": null
	}]
}

Some customers don't have all the information: for example the last customer doesn't have a zipcode.

I noticed that when i convert this avro payload to a TCK record, some nested records have incomplete TCK schema. For example, for the last customer without zipcode, the generated TCK schema linked to the corresonding TCK record do not have a zipcode field. This TCK schema should have a zipcode field since the corresponding avro schema has a zipcode field.

It could be possible to change a little bit the code of AvroToRecord in order to generate complete TCK schema:

• Line 152: add inferSchema(record) when calling recordBuilderFactory.newRecordBuilder

      private void buildArrayField(org.apache.avro.Schema.Field field, Collection<?> value, Record.Builder recordBuilder,
              Entry entry) {
          final org.apache.avro.Schema arraySchema = AvroHelper.getUnionSchema(field.schema());
          final org.apache.avro.Schema arrayInnerType = arraySchema.getElementType();
  
          final Collection<?> objectArray;
          switch (arrayInnerType.getType()) {
          case RECORD:
              objectArray = ((Collection<GenericRecord>) value).stream()
                      .map(record -> avroToRecord(record, arrayInnerType.getFields(), recordBuilderFactory.newRecordBuilder(inferSchema(record))))
                      .collect(Collectors.toList());
              break;
  

• line 188: add inferSchema((GenericRecord) value) when calling recordBuilderFactory.newRecordBuilder

    protected void buildField(org.apache.avro.Schema.Field field, Object value, Record.Builder recordBuilder, Entry entry) {
        String logicalType = field.schema().getProp(AVRO_LOGICAL_TYPE);
        org.apache.avro.Schema.Type fieldType = AvroHelper.getFieldType(field);
        switch (fieldType) {
        case RECORD:
            recordBuilder.withRecord(entry, avroToRecord((GenericRecord) value, ((GenericRecord) value).getSchema().getFields(),
                    recordBuilderFactory.newRecordBuilder(inferSchema((GenericRecord) value))));
            break;