During the development of 6.x Specifically , there are 2 aspects of batch fetch support that can be improved -
# Cache generated SQL AST and friends when possible
# Implement support for different strategies for handling batch-size in different scenarios - ala legacy {{BatchFetchStrategy}} as applied to {{@BatchSize}} for entity and collection fetches was not ported due to time constraints.
In If the interim, {{BatchFetchStrategy}} was deprecated. Which is fine. I think the strategy for entity and collection thing being batch fetching should be fetched has a mix {{DYNAMIC}} and {{PADDED}} single-column key , offer support for loading based on a single SQL ARRAY parameter. Otherwise, fallback to the outcome of “padded” strategy (though potentially partitioning the underlying fetching into smaller chunks depending on {{Dialect#getDefaultBatchLoadSizingStrategy}} .
The current implementation implements For the {{DYNAMIC}} ARRAY strategy .
When {{PADDED}} is used , we should cache the generated SQL AST look at supporting that for multi-load also . When {{DYNAMIC}}, we should maybe offer some limited caching of SQL AST(s). |
|
