...
- Storing only 'books' will allow for removing the 'LIKE' operator from the SQL query, and also allow GET operation (getDataNodes) to use index-only lookups while supporting listspotentially speeding queries in the general case.
- Storing "books[@title='Matilda']" will allow for future optimization of queries to skip the attribute check, for cases where the leaf-condition references the key leaf (as defined in the Yang model), thus skipping the LIKE operator and the attribute check, but only in this case.
A note on fetching descendant nodes
...