Why does an index scan read more pages than exist in an index?

by Radu Gheorghiu   Last Updated January 12, 2018 14:06 PM

I've been trying out running some queries for a testing scenario and I noticed that when doing a full-scan of a nonclustered index, SQL Server reads more pages than actually are in the index.

As you can see from the image below, my query is (a simple one):

    , somemoretext
from tbl

and I have an exact covering nonclustered index for those columns.

enter image description here

In Profiler the number of pages read for this is 22,356 as you can see in the image.

But when looking at the index details, under Fragmentation, the number of pages for this index is 22,278.

enter image description here

What is the cause for this difference in pages read (78 to be precise)? Is the number in the Fragmentation section not reliable or is there some other explanation?

The index is not fragmented (or appears not to be), so there are no more pages than those 22,278 .. or are there? I wasn't able to find an explanation for this online, on Microsoft sites or otherwise.

(This is a controlled environment, no inserts and no deletes occurred on the table while taking screenshots for this question)

I have also pasted the plan here.

Answers 1

You should check what code is submitted to server when you open "Fragmentation" (using SQL Server Profiler)

And you'll see the following code:

select partition_number as PartitionNumber,
       index_type_desc  as IndexType,
       index_depth as Depth,
       avg_fragmentation_in_percent as AverageFragmentation,
       page_count   as Pages,
       avg_page_space_used_in_percent   as AveragePageDensity,
       record_count as Rows,
       ghost_record_count   as GhostRows,
       version_ghost_record_count   as VersionGhostRows,
       min_record_size_in_bytes as MinimumRecordSize,
       max_record_size_in_bytes as MaximumRecordSize,
       avg_record_size_in_bytes as AverageRecordSize,
       forwarded_record_count as ForwardedRecords 
from sys.dm_db_index_physical_stats(7, 773577794, 2, NULL, 'SAMPLED')

Note the mode: it's 'SAMPLED'.

The code returns only 1 row that reflects page_count of the leaf level of the index.

Now change the code to use 'DETAILED' and you'll find missing pages from non-leaf levels.

January 12, 2018 13:59 PM

Related Questions

Full text catalog missing a keyword

Updated August 23, 2017 13:06 PM

MSDTC Issue SQL Server 2008 R2

Updated August 23, 2016 08:04 AM