問題文
An airline's maintenance app keeps millions of Work_Order__c records with a Private org-wide default, and a small number of service leads own most of them. The same filtered query is fast for a technician and slow for a lead. What is the correct assessment of the relationship between the sharing design and query cost?
選択肢
- Sharing is applied only after the database has returned every matching row, so the sharing design can never change how a query is planned.
- Adding share rows always makes queries faster, because each row acts as an index entry that points directly at the record it grants access to.
- Record visibility takes part in how the query is planned, so a filter can be selective for a user who sees few records and not selective for one who sees many; the documented practice is to avoid concentrating ownership of very large numbers of records on one user.
- The sharing design is irrelevant to selectivity because the optimizer measures a filter against the total number of records in the object, and that total is the same no matter who runs the query, which is why one plan is reused for every user in the org, so the two timings must have another cause.