Non-leaf level entries contain only index key columns and the pointer to a lower level page. Included columns only exist in the leaf level entries; they are not carried in the non-leaf level entries.

GilaMonster (1/20/2012)

---

Just one correction:



Non-leaf entries of a nonclustered index contain the index key columns, a pointer to the page below and, if the index has not been declared as unique, it will also contain the clustered index key or RID (depending whether the base table is a heap or clustered index). This ensures that there are never complete duplicate rows in the upper level of a nonclustered index

patrickmcginnis59 (1/20/2012)

---

Whats the downside of having duplicate rows in nonleaf nodes? Aren't they still going to point to the correct child nodes that they're the parent of?

GilaMonster (1/20/2012)

---

Yes, they will, however it can be more work when modifying or deleting the rows. Imagine an index on gender and a row is deleted. That row needs to be located in the nonclustered index to be removed (from the leaf). If all we had at the intermediate levels was the gender, SQL would have to scan half the index to find the row that it needs to remove. If the index rows are unique (from the clustered key or RID), SQL can just navigate straight to the row it need to remove.
http://technet.microsoft.com/en-us/sqlserver/gg508878.aspx (about 24 minutes in)

patrickmcginnis59 (1/20/2012)

---

Thanks! I didn't even consider deletes. I'll have to buy some popcorn and watch the movie!