IndexScan+Filter =(implies)=> Better Index Possible? -- sql server programming

sql server programming : IndexScan+Filter =(implies)=> Better Index Possible?

Ian Boyd
5/26/2004 11:24:36 PM

Part of one of my queries performs
an index scan (generates 1.8M rows), and then
filters that set (down to 61 rows).

The plan text follows.

My question is (because i just don't know) shouldn't an IndexScan followed
by a Filter
operation be able to be converted into an index seek? Or at least not have
it generate
so many candidate rows in the inital scan?

Why not apply the Filter when doing the scan in the first place? You (you
being
the SQL Server engine) would save yourself a lot of work).

You'll notice the Filter operation is performing some calculations on dates,
but if i had a convering index order first by the dates, shouldn't that be
able to be taken care of in the initial scan?

Formatted query plan output of the two offending stages:

Index Scan:
OBJECT: Reservations.IX_Reservations_ConflictsCheckingHelper
WHERE:( Reservations.ReservationGUID <> Reservations.ReservationGUID
AND Reservations.Status <> 'rsCancelled')
AND Reservations.Status <> 'rsNoShow'

RowCount: 1,888,409

Filter
WHERE:
( (Reservations.ArrivalBayGUID=Reservations.ArrivalBayGUID
AND Reservations.ScheduledArrivalDate >= Reservations.ScheduledArrivalDate)
AND Reservations.ScheduledArrivalDate <= dateadd(minute, 15,
Reservations.ScheduledArrivalDate)
)
OR
( (Reservations.DepartureBayGUID=Reservations.ArrivalBayGUID
AND Reservations.ScheduledDepartureDate >=
Reservations.ScheduledArrivalDate)
AND Reservations.ScheduledDepartureDate <= dateadd(minute, 15,
[Reservations].[ScheduledArrivalDate]))
)
OR
( ([Reservations].[ArrivalBayGUID]=[Reservations].[DepartureBayGUID]
AND
[Reservations].[ScheduledArrivalDate]>=[Reservations].[ScheduledDepartureDat
e])
AND [Reservations].[ScheduledArrivalDate]<=dateadd(minute, 15,
[Reservations].[ScheduledDepartureDate]))
)
OR
( ([Reservations].[DepartureBayGUID]=[Reservations].[DepartureBayGUID]
AND
[Reservations].[ScheduledDepartureDate]>=[Reservations].[ScheduledDepartureD
ate])
AND [Reservations].[ScheduledDepartureDate]<=dateadd(minute, 15,
[Reservations].[ScheduledDepartureDate]
)

RowCount: 61

Ian Boyd
5/27/2004 8:22:03 AM

Now i'm kinda confused. The query itself runs faster, and it has a lower
subtree cost (18,000 vs 24)

Table 'Worktable'. Scan count 52, logical reads 58229, physical reads 0,
read-ahead reads 0.
Table 'Reservations'. Scan count 7, logical reads 1717, physical reads 1,
read-ahead reads 78.
Table 'Worktable'. Scan count 56, logical reads 58239, physical reads 0,
read-ahead reads 0.
Table 'Worktable'. Scan count 59, logical reads 48039, physical reads 0,
read-ahead reads 0.
Table 'Worktable'. Scan count 100, logical reads 48121, physical reads 0,
read-ahead reads 0.

But it has a lot more logical reads (i used DBCC DropCleanBuffers before
running query).

(Without drop clean buffers:
Table 'Worktable'. Scan count 52, logical reads 58229, physical reads 0,
read-ahead reads 0.
Table 'Reservations'. Scan count 7, logical reads 1717, physical reads 0,
read-ahead reads 0.
Table 'Worktable'. Scan count 56, logical reads 58239, physical reads 0,
read-ahead reads 0.
Table 'Worktable'. Scan count 59, logical reads 48039, physical reads 0,
read-ahead reads 0.
Table 'Worktable'. Scan count 100, logical reads 48121, physical reads 0,
read-ahead reads 0.
)

So is this really better?

Also, the server wasn't able to perform that conversion on it's own?
Can you explain logically how this is better?

With the OR clauses, i thought it would perform "short-circuit evaluation".
As soon as it finds an or clause that fits the criteria, it can bail out.
Verses having to run all union all conditions?

[quoted text, click to view]

Uri Dimant
5/27/2004 8:34:36 AM

Ian
I suggest you to re-write all 'OR' conditions in the WHERE clause to UNION .
It may cause to query optomizer to use INDEX SEEK.
Also you have to create an appropriate indexes on the table.

For example
SELECT col FROM Table WHERE col1=1 OR col2=2
try to use
SELECT col FROM Table WHERE col1=1
UNION -----ALL
SELECT col FROM Table WHERE col2=2

[quoted text, click to view]

Hugo Kornelis
5/27/2004 3:33:48 PM

[quoted text, click to view]

(snip)
[quoted text, click to view]

Let's assume you have two phone books for New York. One sorted by name,
the other one sorted by address (not commercially available). Now, if You
had to find everybody named White *or* living on 42nd Avenue, would you
browse one book from beginning to end, or would you rather first get the
Whites from one book, then get the people living 42nd Avenue from the
second book, then take some time to remove duplicates?

(snip)

Best, Hugo
--

Got something to say? Click here to post a reply to this thread.

Don't see what you're looking for? Try a search.

View other messages about sql server programming

View Other Months
January 2003
March 2003
April 2003
May 2003
June 2003
July 2003
August 2003
September 2003
October 2003
November 2003
December 2003
January 2004
February 2004
March 2004
April 2004
May 2004
June 2004
July 2004
August 2004
September 2004
October 2004
November 2004
December 2004
January 2005
February 2005
March 2005
April 2005
May 2005
June 2005
July 2005
August 2005
September 2005
October 2005
November 2005
December 2005
January 2006
February 2006
March 2006
April 2006
May 2006
June 2006
July 2006
August 2006
September 2006
October 2006
November 2006
December 2006
January 2007
February 2007
March 2007
April 2007
May 2007
June 2007
July 2007
August 2007
September 2007
October 2007
November 2007
April 2008
August 2008

home · blog · groups

Questions? Comments? Contact the dn