This does not work well, and did not simplify the client-code, at all.
Should an indexer run before or after Mappers are applied?
If we run the indexer before Mappers, the type-conversions won't have been applied to the rows - you end up with raw SQL values.
If we run the indexer after the Mappers, other Mappers may have turned the rows into objects, and the array type-hint in the Indexer::index() method will fail.
Either way, this did not help simplifying client-code - in fact, it just made things more complicated and not very easy to understand.
This does not work well, and did not simplify the client-code, at all.
Should an indexer run before or after Mappers are applied?
If we run the indexer before Mappers, the type-conversions won't have been applied to the rows - you end up with raw SQL values.
If we run the indexer after the Mappers, other Mappers may have turned the rows into objects, and the
array
type-hint in theIndexer::index()
method will fail.Either way, this did not help simplifying client-code - in fact, it just made things more complicated and not very easy to understand.
Closing for now.