If there is a clustered index and the lookup column is not in the key column of the nonclustered index and the RID lookup bookmark is used in a table that does not have a clustered index, it may be because the overhead is too large to cause some queries to directly perform table scan clustered index all the data is in the index (the data page is also part of the index). So it is possible to locate all the columns directly from the clustered index to the index, which means that the data is found but after the nonclustered index is positioned to the index, it is necessary to find the data rows based on the RID or key recorded by the index, which will result in too many I/O data rows for the data page, and the data page IO will be very large , IO may instead exceed the overhead of a direct page scan, so the system chooses to scan the page so that the bookmark lookup is unpredictable and needs to be avoided as much as possible.
07. Index-Nonclustered index (3)-key lookup &rid Lookup
