I ran into a small problem when indexing mailing list archives. Since the e-mail headers are stored in an array of an inner object, the names and values get flattened when indexing. This means that while it's possible to filter on a header like "Message-ID" by using field name "header.value", it's not possible to know if it's that header or another, like "References" or "In-Reply-To" that's actually matching.
Since one probably doesn't want to create fields for all headers, I've added a new river configuration property (headers_to_fields) where one can list headers that should be copied to fields.
A header name like "Message-ID" creates a field with name "header_message_id".
I ran into a small problem when indexing mailing list archives. Since the e-mail headers are stored in an array of an inner object, the names and values get flattened when indexing. This means that while it's possible to filter on a header like "Message-ID" by using field name "header.value", it's not possible to know if it's that header or another, like "References" or "In-Reply-To" that's actually matching.
Since one probably doesn't want to create fields for all headers, I've added a new river configuration property (headers_to_fields) where one can list headers that should be copied to fields.
A header name like "Message-ID" creates a field with name "header_message_id".