Add case-insensitive and order independent field mapping to PutIceberg record converter

The schema matching is currently case-sensitive, and the processor expects the same field ordering in the incoming records and in the target Iceberg schema. To make it less strict the field name matching should be case-insensitive and ordering independent.