|
Significant investigation into the performance of the inventory storage module has generated quite a lot of information about the kinds of SQL queries that are generated from CQL queries and the impact on performance.
It would be good to document this to clarify guidance around what kinds of queries to use in different circumstances and how best to support them in the backend implementations.
Examples of topics:
- == vs. = relation semantics and performance implications
- How index definitions need to match the case sensitivity and accent treatment of the CQL conversion
And possibly to expand upon it to improve the implementation. For example, it appears that a == relation creates a LIKE statement, and I wonder if there could be cases where an exact match could be done with a SQL equals?
|