Adding an additional “order by” column gives me a much worse planHelp optimizing MySQL slow...
Telemetry for feature health
Why didn't Voldemort know what Grindelwald looked like?
Does the Crossbow Expert feat's extra crossbow attack work with the reaction attack from a Hunter ranger's Giant Killer feature?
What is the meaning of "You've never met a graph you didn't like?"
How were servants to the Kaiser of Imperial Germany treated and where may I find more information on them
What (the heck) is a Super Worm Equinox Moon?
What does "tick" mean in this sentence?
Would a primitive species be able to learn English from reading books alone?
Ways of geometrical multiplication
Do you waste sorcery points if you try to apply metamagic to a spell from a scroll but fail to cast it?
How do I tell my boss that I'm quitting in 15 days (a colleague left this week)
Why is the Sun approximated as a black body at ~ 5800 K?
Is there a reason to prefer HFS+ over APFS for disk images in High Sierra and/or Mojave?
Unable to disable Microsoft Store in domain environment
How do you justify more code being written by following clean code practices?
Animation: customize bounce interpolation
Sigmoid with a slope but no asymptotes?
If the only attacker is removed from combat, is a creature still counted as having attacked this turn?
Are Captain Marvel's powers affected by Thanos breaking the Tesseract and claiming the stone?
Should I warn new/interviewing PhD Student that supervisor is terrible?
Why does the Persian emissary display a string of crowned skulls?
Alignment of six matrices
Giving feedback to someone without sounding prejudiced
How many people need to be born every 8 years to sustain population?
Adding an additional “order by” column gives me a much worse plan
Help optimizing MySQL slow queryOptimizing ORDER BY for simple MySQL queryDatabase Implementations of ORDER BY in a Subqueryquery performance gains by removing operator hash match inner joinWhy are these two queries having such different executions?How can I update statistics adding the data of the last day only?left outer join - sort operations in the query plan - any ways of tuning this simple query?Why does changing the declared join column order introduce a sort?When can SARGable predicates be pushed into a CTE or derived table?Why does the location of a join change performance?
in other words, how can I get rid of the sort
operator on the picture below?
the picture above shows the execution plan of the following 2 selects together:
SELECT TOP 1 so.OrgType,
ch.Status,
rcs.DBSstatusDescription,
cid.ApplicationId
FROM tbl_application_crb_initialData cid
INNER JOIN tbl_season_organisationId so
ON cid.OrganisationId = so.OrgId
LEFT JOIN tbl_crbHistory ch
ON cid.ClientReference = ch.ClientReference
LEFT JOIN ref_crbStatus rcs
ON ch.Status = rcs.statusId
ORDER BY cid.DateAdded DESC, ch.DateAdded DESC
SELECT TOP 1 so.OrgType,
ch.Status,
rcs.DBSstatusDescription,
cid.ApplicationId
FROM tbl_application_crb_initialData cid
INNER JOIN tbl_season_organisationId so
ON cid.OrganisationId = so.OrgId
LEFT JOIN tbl_crbHistory ch
ON cid.ClientReference = ch.ClientReference
LEFT JOIN ref_crbStatus rcs
ON ch.Status = rcs.statusId
ORDER BY cid.DateAdded DESC--, ch.DateAdded DESC
the only difference is that on the second query, there is only one column in the order by.
would it make a difference, as I am using top 1
?
I believe all the needed info are on the indexes and table definitions that can be seen on the query plan.
if anything else would help to get rid of that sort
just let me know, tomorrow I will post all the possible info.
sql-server query-performance sql-server-2016 optimization order-by
add a comment |
in other words, how can I get rid of the sort
operator on the picture below?
the picture above shows the execution plan of the following 2 selects together:
SELECT TOP 1 so.OrgType,
ch.Status,
rcs.DBSstatusDescription,
cid.ApplicationId
FROM tbl_application_crb_initialData cid
INNER JOIN tbl_season_organisationId so
ON cid.OrganisationId = so.OrgId
LEFT JOIN tbl_crbHistory ch
ON cid.ClientReference = ch.ClientReference
LEFT JOIN ref_crbStatus rcs
ON ch.Status = rcs.statusId
ORDER BY cid.DateAdded DESC, ch.DateAdded DESC
SELECT TOP 1 so.OrgType,
ch.Status,
rcs.DBSstatusDescription,
cid.ApplicationId
FROM tbl_application_crb_initialData cid
INNER JOIN tbl_season_organisationId so
ON cid.OrganisationId = so.OrgId
LEFT JOIN tbl_crbHistory ch
ON cid.ClientReference = ch.ClientReference
LEFT JOIN ref_crbStatus rcs
ON ch.Status = rcs.statusId
ORDER BY cid.DateAdded DESC--, ch.DateAdded DESC
the only difference is that on the second query, there is only one column in the order by.
would it make a difference, as I am using top 1
?
I believe all the needed info are on the indexes and table definitions that can be seen on the query plan.
if anything else would help to get rid of that sort
just let me know, tomorrow I will post all the possible info.
sql-server query-performance sql-server-2016 optimization order-by
Those cost % are just estimates and can actually end up being WAY, WAY off. Is the top query actually noticeably slower?
– Aaron Bertrand♦
Mar 12 at 18:34
1
@AaronBertrand -CPU time 756, Elapsed Time 222, Reads 2357
vsCPU time 0, Elapsed Time 0, Reads 10
from the stats in the plan - most of the tables are pretty small though. Table cardinalities11
,19
,67,591
,232,528
– Martin Smith
Mar 12 at 19:33
1
@MartinSmith Thanks, I didn’t look at the plan (mobile), just try my best to make sure drive-by readers don’t put too much weight into those percentages. Sometimes they’re useful, sometimes they’re extremely misleading.
– Aaron Bertrand♦
Mar 12 at 20:57
add a comment |
in other words, how can I get rid of the sort
operator on the picture below?
the picture above shows the execution plan of the following 2 selects together:
SELECT TOP 1 so.OrgType,
ch.Status,
rcs.DBSstatusDescription,
cid.ApplicationId
FROM tbl_application_crb_initialData cid
INNER JOIN tbl_season_organisationId so
ON cid.OrganisationId = so.OrgId
LEFT JOIN tbl_crbHistory ch
ON cid.ClientReference = ch.ClientReference
LEFT JOIN ref_crbStatus rcs
ON ch.Status = rcs.statusId
ORDER BY cid.DateAdded DESC, ch.DateAdded DESC
SELECT TOP 1 so.OrgType,
ch.Status,
rcs.DBSstatusDescription,
cid.ApplicationId
FROM tbl_application_crb_initialData cid
INNER JOIN tbl_season_organisationId so
ON cid.OrganisationId = so.OrgId
LEFT JOIN tbl_crbHistory ch
ON cid.ClientReference = ch.ClientReference
LEFT JOIN ref_crbStatus rcs
ON ch.Status = rcs.statusId
ORDER BY cid.DateAdded DESC--, ch.DateAdded DESC
the only difference is that on the second query, there is only one column in the order by.
would it make a difference, as I am using top 1
?
I believe all the needed info are on the indexes and table definitions that can be seen on the query plan.
if anything else would help to get rid of that sort
just let me know, tomorrow I will post all the possible info.
sql-server query-performance sql-server-2016 optimization order-by
in other words, how can I get rid of the sort
operator on the picture below?
the picture above shows the execution plan of the following 2 selects together:
SELECT TOP 1 so.OrgType,
ch.Status,
rcs.DBSstatusDescription,
cid.ApplicationId
FROM tbl_application_crb_initialData cid
INNER JOIN tbl_season_organisationId so
ON cid.OrganisationId = so.OrgId
LEFT JOIN tbl_crbHistory ch
ON cid.ClientReference = ch.ClientReference
LEFT JOIN ref_crbStatus rcs
ON ch.Status = rcs.statusId
ORDER BY cid.DateAdded DESC, ch.DateAdded DESC
SELECT TOP 1 so.OrgType,
ch.Status,
rcs.DBSstatusDescription,
cid.ApplicationId
FROM tbl_application_crb_initialData cid
INNER JOIN tbl_season_organisationId so
ON cid.OrganisationId = so.OrgId
LEFT JOIN tbl_crbHistory ch
ON cid.ClientReference = ch.ClientReference
LEFT JOIN ref_crbStatus rcs
ON ch.Status = rcs.statusId
ORDER BY cid.DateAdded DESC--, ch.DateAdded DESC
the only difference is that on the second query, there is only one column in the order by.
would it make a difference, as I am using top 1
?
I believe all the needed info are on the indexes and table definitions that can be seen on the query plan.
if anything else would help to get rid of that sort
just let me know, tomorrow I will post all the possible info.
sql-server query-performance sql-server-2016 optimization order-by
sql-server query-performance sql-server-2016 optimization order-by
edited Mar 12 at 22:22
Martin Smith
63.9k10172256
63.9k10172256
asked Mar 12 at 18:26
marcello miorellimarcello miorelli
5,9312162141
5,9312162141
Those cost % are just estimates and can actually end up being WAY, WAY off. Is the top query actually noticeably slower?
– Aaron Bertrand♦
Mar 12 at 18:34
1
@AaronBertrand -CPU time 756, Elapsed Time 222, Reads 2357
vsCPU time 0, Elapsed Time 0, Reads 10
from the stats in the plan - most of the tables are pretty small though. Table cardinalities11
,19
,67,591
,232,528
– Martin Smith
Mar 12 at 19:33
1
@MartinSmith Thanks, I didn’t look at the plan (mobile), just try my best to make sure drive-by readers don’t put too much weight into those percentages. Sometimes they’re useful, sometimes they’re extremely misleading.
– Aaron Bertrand♦
Mar 12 at 20:57
add a comment |
Those cost % are just estimates and can actually end up being WAY, WAY off. Is the top query actually noticeably slower?
– Aaron Bertrand♦
Mar 12 at 18:34
1
@AaronBertrand -CPU time 756, Elapsed Time 222, Reads 2357
vsCPU time 0, Elapsed Time 0, Reads 10
from the stats in the plan - most of the tables are pretty small though. Table cardinalities11
,19
,67,591
,232,528
– Martin Smith
Mar 12 at 19:33
1
@MartinSmith Thanks, I didn’t look at the plan (mobile), just try my best to make sure drive-by readers don’t put too much weight into those percentages. Sometimes they’re useful, sometimes they’re extremely misleading.
– Aaron Bertrand♦
Mar 12 at 20:57
Those cost % are just estimates and can actually end up being WAY, WAY off. Is the top query actually noticeably slower?
– Aaron Bertrand♦
Mar 12 at 18:34
Those cost % are just estimates and can actually end up being WAY, WAY off. Is the top query actually noticeably slower?
– Aaron Bertrand♦
Mar 12 at 18:34
1
1
@AaronBertrand -
CPU time 756, Elapsed Time 222, Reads 2357
vs CPU time 0, Elapsed Time 0, Reads 10
from the stats in the plan - most of the tables are pretty small though. Table cardinalities 11
, 19
, 67,591
, 232,528
– Martin Smith
Mar 12 at 19:33
@AaronBertrand -
CPU time 756, Elapsed Time 222, Reads 2357
vs CPU time 0, Elapsed Time 0, Reads 10
from the stats in the plan - most of the tables are pretty small though. Table cardinalities 11
, 19
, 67,591
, 232,528
– Martin Smith
Mar 12 at 19:33
1
1
@MartinSmith Thanks, I didn’t look at the plan (mobile), just try my best to make sure drive-by readers don’t put too much weight into those percentages. Sometimes they’re useful, sometimes they’re extremely misleading.
– Aaron Bertrand♦
Mar 12 at 20:57
@MartinSmith Thanks, I didn’t look at the plan (mobile), just try my best to make sure drive-by readers don’t put too much weight into those percentages. Sometimes they’re useful, sometimes they’re extremely misleading.
– Aaron Bertrand♦
Mar 12 at 20:57
add a comment |
1 Answer
1
active
oldest
votes
Your question is missing a lot of detail but I can reproduce something similar.
Setup
CREATE TABLE T1(X INT PRIMARY KEY, Y INT INDEX IX)
CREATE TABLE T2(X INT, Y INT , PRIMARY KEY(X, Y))
INSERT INTO T2
OUTPUT INSERTED.* INTO T1
SELECT TOP 1000000 ROW_NUMBER() OVER (ORDER BY @@SPID), ROW_NUMBER() OVER (ORDER BY @@SPID)
FROM sys.all_objects o1, sys.all_objects o2;
Query 1
SELECT TOP 1 T1.Y AS T1Y, T2.Y AS T2Y
FROM T1 JOIN T2 ON T1.X = T2.X
ORDER BY T1.Y;
Query 2
SELECT TOP 1 T1.Y AS T1Y, T2.Y AS T2Y
FROM T1 JOIN T2 ON T1.X = T2.X
ORDER BY T1.Y, T2.Y
Query 3
WITH T AS
(
SELECT TOP 1 WITH TIES T1.Y AS T1Y, T2.Y AS T2Y
FROM T1 JOIN T2 ON T1.X = T2.X
ORDER BY T1.Y
)
SELECT TOP 1 *
FROM T
ORDER BY T2Y
Query 1 just picks off the TOP 1
from the index in the desired sort order and does the needed joins on the other table for that row. If the join is successful it stops there otherwise it tries the next one in index order until it finds a row that matches or runs out of rows.
Query 2 When adding the new sort column this plan is no longer valid as there could be multiple matches tied with the TOP 1
value and SQL Server decides to join the whole lot and then get the TOP 1
from that.
Query 3 This encourages SQL Server to stick with the first strategy and then just does a TOP 1 Sort
on any rows tied with the same value for the first sort key.
For my example data Query 3 works out better than Query 2 but if you have many duplicates tied for the value of the first sort key your milage may differ.
You can try this rewrite and see how it fares
WITH T
AS (SELECT TOP 1 WITH TIES so.OrgType,
ch.Status,
rcs.DBSstatusDescription,
cid.ApplicationId,
ch.DateAdded AS chDateAdded
FROM tbl_application_crb_initialData cid
INNER JOIN tbl_season_organisationId so
ON cid.OrganisationId = so.OrgId
LEFT JOIN tbl_crbHistory ch
ON cid.ClientReference = ch.ClientReference
LEFT JOIN ref_crbStatus rcs
ON ch.Status = rcs.statusId
ORDER BY cid.DateAdded DESC)
SELECT TOP 1 OrgType,
Status,
DBSstatusDescription,
ApplicationId
FROM T
ORDER BY chDateAdded DESC
1
A+ for the first time I've seen someone useTOP WITH TIES
for a good reason.
– Erik Darling
Mar 13 at 2:01
add a comment |
Your Answer
StackExchange.ready(function() {
var channelOptions = {
tags: "".split(" "),
id: "182"
};
initTagRenderer("".split(" "), "".split(" "), channelOptions);
StackExchange.using("externalEditor", function() {
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled) {
StackExchange.using("snippets", function() {
createEditor();
});
}
else {
createEditor();
}
});
function createEditor() {
StackExchange.prepareEditor({
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: false,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: null,
bindNavPrevention: true,
postfix: "",
imageUploader: {
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
},
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
});
}
});
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdba.stackexchange.com%2fquestions%2f231950%2fadding-an-additional-order-by-column-gives-me-a-much-worse-plan%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
1 Answer
1
active
oldest
votes
1 Answer
1
active
oldest
votes
active
oldest
votes
active
oldest
votes
Your question is missing a lot of detail but I can reproduce something similar.
Setup
CREATE TABLE T1(X INT PRIMARY KEY, Y INT INDEX IX)
CREATE TABLE T2(X INT, Y INT , PRIMARY KEY(X, Y))
INSERT INTO T2
OUTPUT INSERTED.* INTO T1
SELECT TOP 1000000 ROW_NUMBER() OVER (ORDER BY @@SPID), ROW_NUMBER() OVER (ORDER BY @@SPID)
FROM sys.all_objects o1, sys.all_objects o2;
Query 1
SELECT TOP 1 T1.Y AS T1Y, T2.Y AS T2Y
FROM T1 JOIN T2 ON T1.X = T2.X
ORDER BY T1.Y;
Query 2
SELECT TOP 1 T1.Y AS T1Y, T2.Y AS T2Y
FROM T1 JOIN T2 ON T1.X = T2.X
ORDER BY T1.Y, T2.Y
Query 3
WITH T AS
(
SELECT TOP 1 WITH TIES T1.Y AS T1Y, T2.Y AS T2Y
FROM T1 JOIN T2 ON T1.X = T2.X
ORDER BY T1.Y
)
SELECT TOP 1 *
FROM T
ORDER BY T2Y
Query 1 just picks off the TOP 1
from the index in the desired sort order and does the needed joins on the other table for that row. If the join is successful it stops there otherwise it tries the next one in index order until it finds a row that matches or runs out of rows.
Query 2 When adding the new sort column this plan is no longer valid as there could be multiple matches tied with the TOP 1
value and SQL Server decides to join the whole lot and then get the TOP 1
from that.
Query 3 This encourages SQL Server to stick with the first strategy and then just does a TOP 1 Sort
on any rows tied with the same value for the first sort key.
For my example data Query 3 works out better than Query 2 but if you have many duplicates tied for the value of the first sort key your milage may differ.
You can try this rewrite and see how it fares
WITH T
AS (SELECT TOP 1 WITH TIES so.OrgType,
ch.Status,
rcs.DBSstatusDescription,
cid.ApplicationId,
ch.DateAdded AS chDateAdded
FROM tbl_application_crb_initialData cid
INNER JOIN tbl_season_organisationId so
ON cid.OrganisationId = so.OrgId
LEFT JOIN tbl_crbHistory ch
ON cid.ClientReference = ch.ClientReference
LEFT JOIN ref_crbStatus rcs
ON ch.Status = rcs.statusId
ORDER BY cid.DateAdded DESC)
SELECT TOP 1 OrgType,
Status,
DBSstatusDescription,
ApplicationId
FROM T
ORDER BY chDateAdded DESC
1
A+ for the first time I've seen someone useTOP WITH TIES
for a good reason.
– Erik Darling
Mar 13 at 2:01
add a comment |
Your question is missing a lot of detail but I can reproduce something similar.
Setup
CREATE TABLE T1(X INT PRIMARY KEY, Y INT INDEX IX)
CREATE TABLE T2(X INT, Y INT , PRIMARY KEY(X, Y))
INSERT INTO T2
OUTPUT INSERTED.* INTO T1
SELECT TOP 1000000 ROW_NUMBER() OVER (ORDER BY @@SPID), ROW_NUMBER() OVER (ORDER BY @@SPID)
FROM sys.all_objects o1, sys.all_objects o2;
Query 1
SELECT TOP 1 T1.Y AS T1Y, T2.Y AS T2Y
FROM T1 JOIN T2 ON T1.X = T2.X
ORDER BY T1.Y;
Query 2
SELECT TOP 1 T1.Y AS T1Y, T2.Y AS T2Y
FROM T1 JOIN T2 ON T1.X = T2.X
ORDER BY T1.Y, T2.Y
Query 3
WITH T AS
(
SELECT TOP 1 WITH TIES T1.Y AS T1Y, T2.Y AS T2Y
FROM T1 JOIN T2 ON T1.X = T2.X
ORDER BY T1.Y
)
SELECT TOP 1 *
FROM T
ORDER BY T2Y
Query 1 just picks off the TOP 1
from the index in the desired sort order and does the needed joins on the other table for that row. If the join is successful it stops there otherwise it tries the next one in index order until it finds a row that matches or runs out of rows.
Query 2 When adding the new sort column this plan is no longer valid as there could be multiple matches tied with the TOP 1
value and SQL Server decides to join the whole lot and then get the TOP 1
from that.
Query 3 This encourages SQL Server to stick with the first strategy and then just does a TOP 1 Sort
on any rows tied with the same value for the first sort key.
For my example data Query 3 works out better than Query 2 but if you have many duplicates tied for the value of the first sort key your milage may differ.
You can try this rewrite and see how it fares
WITH T
AS (SELECT TOP 1 WITH TIES so.OrgType,
ch.Status,
rcs.DBSstatusDescription,
cid.ApplicationId,
ch.DateAdded AS chDateAdded
FROM tbl_application_crb_initialData cid
INNER JOIN tbl_season_organisationId so
ON cid.OrganisationId = so.OrgId
LEFT JOIN tbl_crbHistory ch
ON cid.ClientReference = ch.ClientReference
LEFT JOIN ref_crbStatus rcs
ON ch.Status = rcs.statusId
ORDER BY cid.DateAdded DESC)
SELECT TOP 1 OrgType,
Status,
DBSstatusDescription,
ApplicationId
FROM T
ORDER BY chDateAdded DESC
1
A+ for the first time I've seen someone useTOP WITH TIES
for a good reason.
– Erik Darling
Mar 13 at 2:01
add a comment |
Your question is missing a lot of detail but I can reproduce something similar.
Setup
CREATE TABLE T1(X INT PRIMARY KEY, Y INT INDEX IX)
CREATE TABLE T2(X INT, Y INT , PRIMARY KEY(X, Y))
INSERT INTO T2
OUTPUT INSERTED.* INTO T1
SELECT TOP 1000000 ROW_NUMBER() OVER (ORDER BY @@SPID), ROW_NUMBER() OVER (ORDER BY @@SPID)
FROM sys.all_objects o1, sys.all_objects o2;
Query 1
SELECT TOP 1 T1.Y AS T1Y, T2.Y AS T2Y
FROM T1 JOIN T2 ON T1.X = T2.X
ORDER BY T1.Y;
Query 2
SELECT TOP 1 T1.Y AS T1Y, T2.Y AS T2Y
FROM T1 JOIN T2 ON T1.X = T2.X
ORDER BY T1.Y, T2.Y
Query 3
WITH T AS
(
SELECT TOP 1 WITH TIES T1.Y AS T1Y, T2.Y AS T2Y
FROM T1 JOIN T2 ON T1.X = T2.X
ORDER BY T1.Y
)
SELECT TOP 1 *
FROM T
ORDER BY T2Y
Query 1 just picks off the TOP 1
from the index in the desired sort order and does the needed joins on the other table for that row. If the join is successful it stops there otherwise it tries the next one in index order until it finds a row that matches or runs out of rows.
Query 2 When adding the new sort column this plan is no longer valid as there could be multiple matches tied with the TOP 1
value and SQL Server decides to join the whole lot and then get the TOP 1
from that.
Query 3 This encourages SQL Server to stick with the first strategy and then just does a TOP 1 Sort
on any rows tied with the same value for the first sort key.
For my example data Query 3 works out better than Query 2 but if you have many duplicates tied for the value of the first sort key your milage may differ.
You can try this rewrite and see how it fares
WITH T
AS (SELECT TOP 1 WITH TIES so.OrgType,
ch.Status,
rcs.DBSstatusDescription,
cid.ApplicationId,
ch.DateAdded AS chDateAdded
FROM tbl_application_crb_initialData cid
INNER JOIN tbl_season_organisationId so
ON cid.OrganisationId = so.OrgId
LEFT JOIN tbl_crbHistory ch
ON cid.ClientReference = ch.ClientReference
LEFT JOIN ref_crbStatus rcs
ON ch.Status = rcs.statusId
ORDER BY cid.DateAdded DESC)
SELECT TOP 1 OrgType,
Status,
DBSstatusDescription,
ApplicationId
FROM T
ORDER BY chDateAdded DESC
Your question is missing a lot of detail but I can reproduce something similar.
Setup
CREATE TABLE T1(X INT PRIMARY KEY, Y INT INDEX IX)
CREATE TABLE T2(X INT, Y INT , PRIMARY KEY(X, Y))
INSERT INTO T2
OUTPUT INSERTED.* INTO T1
SELECT TOP 1000000 ROW_NUMBER() OVER (ORDER BY @@SPID), ROW_NUMBER() OVER (ORDER BY @@SPID)
FROM sys.all_objects o1, sys.all_objects o2;
Query 1
SELECT TOP 1 T1.Y AS T1Y, T2.Y AS T2Y
FROM T1 JOIN T2 ON T1.X = T2.X
ORDER BY T1.Y;
Query 2
SELECT TOP 1 T1.Y AS T1Y, T2.Y AS T2Y
FROM T1 JOIN T2 ON T1.X = T2.X
ORDER BY T1.Y, T2.Y
Query 3
WITH T AS
(
SELECT TOP 1 WITH TIES T1.Y AS T1Y, T2.Y AS T2Y
FROM T1 JOIN T2 ON T1.X = T2.X
ORDER BY T1.Y
)
SELECT TOP 1 *
FROM T
ORDER BY T2Y
Query 1 just picks off the TOP 1
from the index in the desired sort order and does the needed joins on the other table for that row. If the join is successful it stops there otherwise it tries the next one in index order until it finds a row that matches or runs out of rows.
Query 2 When adding the new sort column this plan is no longer valid as there could be multiple matches tied with the TOP 1
value and SQL Server decides to join the whole lot and then get the TOP 1
from that.
Query 3 This encourages SQL Server to stick with the first strategy and then just does a TOP 1 Sort
on any rows tied with the same value for the first sort key.
For my example data Query 3 works out better than Query 2 but if you have many duplicates tied for the value of the first sort key your milage may differ.
You can try this rewrite and see how it fares
WITH T
AS (SELECT TOP 1 WITH TIES so.OrgType,
ch.Status,
rcs.DBSstatusDescription,
cid.ApplicationId,
ch.DateAdded AS chDateAdded
FROM tbl_application_crb_initialData cid
INNER JOIN tbl_season_organisationId so
ON cid.OrganisationId = so.OrgId
LEFT JOIN tbl_crbHistory ch
ON cid.ClientReference = ch.ClientReference
LEFT JOIN ref_crbStatus rcs
ON ch.Status = rcs.statusId
ORDER BY cid.DateAdded DESC)
SELECT TOP 1 OrgType,
Status,
DBSstatusDescription,
ApplicationId
FROM T
ORDER BY chDateAdded DESC
edited Mar 12 at 20:54
answered Mar 12 at 19:11
Martin SmithMartin Smith
63.9k10172256
63.9k10172256
1
A+ for the first time I've seen someone useTOP WITH TIES
for a good reason.
– Erik Darling
Mar 13 at 2:01
add a comment |
1
A+ for the first time I've seen someone useTOP WITH TIES
for a good reason.
– Erik Darling
Mar 13 at 2:01
1
1
A+ for the first time I've seen someone use
TOP WITH TIES
for a good reason.– Erik Darling
Mar 13 at 2:01
A+ for the first time I've seen someone use
TOP WITH TIES
for a good reason.– Erik Darling
Mar 13 at 2:01
add a comment |
Thanks for contributing an answer to Database Administrators Stack Exchange!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdba.stackexchange.com%2fquestions%2f231950%2fadding-an-additional-order-by-column-gives-me-a-much-worse-plan%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Those cost % are just estimates and can actually end up being WAY, WAY off. Is the top query actually noticeably slower?
– Aaron Bertrand♦
Mar 12 at 18:34
1
@AaronBertrand -
CPU time 756, Elapsed Time 222, Reads 2357
vsCPU time 0, Elapsed Time 0, Reads 10
from the stats in the plan - most of the tables are pretty small though. Table cardinalities11
,19
,67,591
,232,528
– Martin Smith
Mar 12 at 19:33
1
@MartinSmith Thanks, I didn’t look at the plan (mobile), just try my best to make sure drive-by readers don’t put too much weight into those percentages. Sometimes they’re useful, sometimes they’re extremely misleading.
– Aaron Bertrand♦
Mar 12 at 20:57