I am looking to rank search results in a SQL query - sql-server

The query:
SELECT TOP 1000 t.columns
FROM dbo.teams as t
RIGHT JOIN fnParseString('Nugget,Tulsa',',') ps ON t.team_name LIKE '%'+ps.string+'%' OR t.nickname LIKE '%'+ps.string+'%'
This does return the results I want, but but the ordering isn't useful.
I may add more columns to the search.
How do I rank the rows based on accuracy and number of matched terms? I know it only needs to match one of the terms to be selected, is there a way to then count the number of terms that match the columns.
I see Need help with SQL for ranking search results but will the subselect work for an arbitrary number of search tokens

If you are open to some dynamic SQL
Declare #SearchFor varchar(max) ='Daily,Production,default' -- Any comma delim string
Declare #SearchFrom varchar(150) ='OD' -- table or even a join statemtn
Declare #SearchExpr varchar(150) ='[OD-Title]' -- Any field or even expression
Declare #ReturnCols varchar(150) ='[OD-Nr],[OD-Title]' -- Any field(s) even with alias
Set #SearchFor = 'Sign(CharIndex('''+Replace(Replace(Replace(#SearchFor,' , ',','),', ',''),',',''','+#SearchExpr+'))+Sign(CharIndex(''')+''','+#SearchExpr+'))'
Declare #SQL nvarchar(Max) = 'Select * from (Select Distinct'+#ReturnCols+',Hits='+#SearchFor+' From '+#SearchFrom + ') A Where Hits>0 Order by Hits Desc'
Exec(#SQL)
Returns
OD-Nr OD-Title Hits
3 Daily Production Summary 2
6 Default Settings 1

Non-dynamic option, for contrast. You'd have to add to the order by and the join as you add more columns to check.
SELECT TOP 1000 t.columns
FROM dbo.teams as t
RIGHT JOIN fnParseString('Nugget,Tulsa',',') ps
ON t.team_name LIKE '%'+ps.string+'%'
OR t.nickname LIKE '%'+ps.string+'%'
order by
case when t.team_name LIKE '%'+ps.string+'%' then 1 else 0 +
case when t.nickname LIKE '%'+ps.string+'%' then 1 else 0
desc

Related

MS-SQL getting case specific values

We have a table LogicalTableSharing as follows:
For a specific requirement, we need to take PhysicalCompany of each TableCode into a variable.
We tried a case-based query as follows:
declare #tablecode varchar(50)
declare #inputcompany varchar(50)
declare #query nvarchar(2500)
set #inputcompany= 91
set #query = '
select '''+#inputcompany+''' AS inputcompany,
CASE WHEN lts.TableCode = ''tsctm005'' THEN lts.PhysicalCompany ELSE NULL END as tsctm005_company,
CASE WHEN lts.TableCode = ''tccom000'' THEN lts.PhysicalCompany ELSE NULL END as tccom000_company
from LogicalTableSharing lts
where lts.LogicalCompany = '''+#inputcompany+'''
'
EXEC sp_executesql #query
which obviously gives the result as
The desired output is
What is the right approach?
Try subqueries in a FROM-less SELECT. For performance you want an index on (logicalcompany, tablecode) (or the other way round depending on which is more selective).
SELECT #inputcompany inputcompany,
(SELECT TOP 1
physicalcompany
WHERE logicalcompany = #inputcompany
AND tablecode = 'tsctm005'
ORDER BY <some criteria>) tsctm005_company,
(SELECT TOP 1
physicalcompany
WHERE logicalcompany = #inputcompany
AND tablecode = 'tccom000'
ORDER BY <some criteria>) tccom000_company;
You should find <some criteria> to order by in case of multiple possible rows to decide which one takes precedence. Unless you just want a random one possibly each time you run the query another one, that is.

Dynamic SQL - How to use a value from a result as a column name?

I am working on a dynamic SQL query for a MsSQL stored procedure.
There is a table search.ProfileFields that contains the actual column names in a table I need to query.
My goal is to have the SQL select the specific column in the table, dynamically from its parent query..
A little confusing, heres an example:
DECLARE #sql nvarchar(max)
SELECT #sql = '
SELECT
pfs.SectionID,
pfs.SectionName,
pfs.Link,
(
SELECT
pf.FieldID,
pf.FieldTitle,
pf.FieldSQL,
pf.Restricted,
pf.Optional,
(
SELECT
pf.FieldSQL
FROM
Resources.emp.EmployeeComplete as e
WHERE
e.QID = #QID
) as Value
FROM
search.ProfileFields as pf
WHERE
pf.SectionID = pfs.SectionID
ORDER BY
pf.[Order]
FOR XML PATH (''field''), ELEMENTS, TYPE, ROOT (''fields'')
)
FROM
search.ProfileFieldSections as pfs
WHERE
pfs.Status = 1
FOR XML PATH (''data''), ELEMENTS, TYPE, ROOT (''root'')'
PRINT #sql
EXECUTE sp_executesql #sql, N'#QID varchar(10)', #QID = #QID
In the inner most select. I am querying pf.FieldSQL. I am looking for the actual value that was received by the parent select.
search.ProfileFields has a column called FieldSQL with a few results such as Name, Age, Location.
That is what I am trying to get my inner most select to do.
SELECT Name FROM ... - Name in this case comes from the value of pf.FieldSQL.
How can I go about querying a dynamic column name in this situation?
Have a look at this answer for a couple of suggestions. If your table definition is complex or changes occasionally you probably should use pivot. Here's one that might work for you, so long as column names in the FieldSQL column are well defined, there are not too many of them, and they don't ever change or get added to:
DECLARE #sql nvarchar(max)
SELECT #sql = '
SELECT
pfs.SectionID,
pfs.SectionName,
pfs.Link,
(
SELECT
pf.FieldID,
pf.FieldTitle,
pf.FieldSQL,
pf.Restricted,
pf.Optional,
(
SELECT case pf.FieldSQL
when 'Name' then e.Name
when 'DOB' then convert(nvarchar(10), e.DOB, 126)
-- ... etc.
-- NOTE: may need to aggregated depending on uniqueness of QID:
-- when 'Name' then min(e.Name)
-- when 'DOB' then convert(nvarchar(10), min(e.DOB), 126)
end
FROM
Resources.emp.EmployeeComplete as e
WHERE
e.QID = #QID
) as Value
FROM
search.ProfileFields as pf
WHERE
pf.SectionID = pfs.SectionID
ORDER BY
pf.[Order]
FOR XML PATH (''field''), ELEMENTS, TYPE, ROOT (''fields'')
)
FROM
search.ProfileFieldSections as pfs
WHERE
pfs.Status = 1
FOR XML PATH (''data''), ELEMENTS, TYPE, ROOT (''root'')'
PRINT #sql
EXECUTE sp_executesql #sql, N'#QID varchar(10)', #QID = #QID
Take a look at "PIVOT" operators here PIVOT
Thisshould give you some ideas how to use them.

SQL Server searching for text in a column

I'm confused over what to use?
Basically I need to have a search string that can search a single column for the occurrences of multiple phrases, each input phrase is separated by a space.
So input from user would be like:
"Phrase1 Phrase2 ... PhraseX" (number of phrases can 0 to unknown!, but say < 6)
I need to search with the logic:
Where 'Phrase1%' **AND** 'Phrase2%' **AND** ... 'PhraseX%'
.. etc... so all phrases need to be found.
Always logical AND
SO speed, performance taken in to account, Do I use:
Lot's of
Like 'Phrase1%' and like 'Phrase2%' and like ... 'PhraseX%' ?
or use
patindex('Phrase1%' , column) > 0 AND patindex('Phrase2%' , column) > 0
AND ... patindex('PhraseX%' , column)
or use
add a Full Text search Index,
The use:
Where Contatins(Column, 'Phrase1*') AND Contatins(Column, 'Phrase2*') AND ... Contatins(Column, 'PhraseX*')
Or
????
Almost too many options, which is why I'm asking, what would be the most efficient way of doing this...
Your wisdom is appreciated...
If you are searching for AND, then the correct wildcard search would be:
Like '%Phrase1%' and like '%Phrase2%' and like ... '%PhraseX%'
There is no reason to use patindex() here, because like is sufficient and well optimized. Well optimized, but this case cannot be made efficient. It is going to require a full table scan. And, if the text field is really, really big (I mean at least thousands or tens of thousands of characters), performance will not be good.
The solution is full text search. You would phrase this as:
where CONTAINS(column, 'Phrase1 AND phrase2 AND . . . ');
The only issue here would be when the "phrases" (which seem to be words) you are looking for are stop words.
In conclusion, if you have more than a few thousand rows or the text field you are searching has more than a few thousand characters, then use the full text option. This is just for guidance. If you are searching through a reference table with 100 rows and looking in the description field that has up to 100 characters, then the like method should be fine.
Personally I like this solution -
DECLARE #temp TABLE (title NVARCHAR(50))
INSERT INTO #temp (title)
VALUES ('Phrase1 33'), ('test Phrase2'), ('blank')
SELECT t.*
FROM #temp t
WHERE EXISTS(
SELECT 1
FROM (
VALUES ('Phrase1'), ('Phrase2'), ('PhraseX')
) c(t)
WHERE title LIKE '%' + t + '%'
)
This should ideally be done with the help of Full text search as mentioned above.
BUT,
If you don't have full text configured for your DB, here is a performance intensive solution for doing a prioritized string search. Note : this returns rows for a partial/complete combination of input words(rows containing one or more words of the search string in any order):-
-- table to search in
drop table if exists dbo.myTable;
go
CREATE TABLE dbo.myTable
(
myTableId int NOT NULL IDENTITY (1, 1),
code varchar(200) NOT NULL,
description varchar(200) NOT NULL -- this column contains the values we are going to search in
) ON [PRIMARY]
GO
-- function to split space separated search string into individual words
drop function if exists [dbo].[fnSplit];
go
CREATE FUNCTION [dbo].[fnSplit] (#StringInput nvarchar(max),
#Delimiter nvarchar(1))
RETURNS #OutputTable TABLE (
id nvarchar(1000)
)
AS
BEGIN
DECLARE #String nvarchar(100);
WHILE LEN(#StringInput) > 0
BEGIN
SET #String = LEFT(#StringInput, ISNULL(NULLIF(CHARINDEX(#Delimiter, #StringInput) - 1, -1),
LEN(#StringInput)));
SET #StringInput = SUBSTRING(#StringInput, ISNULL(NULLIF(CHARINDEX
(
#Delimiter, #StringInput
),
0
), LEN
(
#StringInput)
)
+ 1, LEN(#StringInput));
INSERT INTO #OutputTable (id)
VALUES (#String);
END;
RETURN;
END;
GO
-- this is the search script which can be optionally converted to a stored procedure /function
declare #search varchar(max) = 'infection upper acute genito'; -- enter your search string here
-- the searched string above should give rows containing the following
-- infection in upper side with acute genitointestinal tract
-- acute infection in upper teeth
-- acute genitointestinal pain
if (len(trim(#search)) = 0) -- if search string is empty, just return records ordered alphabetically
begin
select 1 as Priority ,myTableid, code, Description from myTable order by Description
return;
end
declare #splitTable Table(
wordRank int Identity(1,1), -- individual words are assinged priority order (in order of occurence/position)
word varchar(200)
)
declare #nonWordTable Table( -- table to trim out auxiliary verbs, prepositions etc. from the search
id varchar(200)
)
insert into #nonWordTable values
('of'),
('with'),
('at'),
('in'),
('for'),
('on'),
('by'),
('like'),
('up'),
('off'),
('near'),
('is'),
('are'),
(','),
(':'),
(';')
insert into #splitTable
select id from dbo.fnSplit(#search,' '); -- this function gives you a table with rows containing all the space separated words of the search like in this e.g., the output will be -
-- id
-------------
-- infection
-- upper
-- acute
-- genito
delete s from #splitTable s join #nonWordTable n on s.word = n.id; -- trimming out non-words here
declare #countOfSearchStrings int = (select count(word) from #splitTable); -- count of space separated words for search
declare #highestPriority int = POWER(#countOfSearchStrings,3);
with plainMatches as
(
select myTableid, #highestPriority as Priority from myTable where Description like #search -- exact matches have highest priority
union
select myTableid, #highestPriority-1 as Priority from myTable where Description like #search + '%' -- then with something at the end
union
select myTableid, #highestPriority-2 as Priority from myTable where Description like '%' + #search -- then with something at the beginning
union
select myTableid, #highestPriority-3 as Priority from myTable where Description like '%' + #search + '%' -- then if the word falls somewhere in between
),
splitWordMatches as( -- give each searched word a rank based on its position in the searched string
-- and calculate its char index in the field to search
select myTable.myTableid, (#countOfSearchStrings - s.wordRank) as Priority, s.word,
wordIndex = CHARINDEX(s.word, myTable.Description) from myTable join #splitTable s on myTable.Description like '%'+ s.word + '%'
-- and not exists(select myTableid from plainMatches p where p.myTableId = myTable.myTableId) -- need not look into rows that have already been found in plainmatches as they are highest ranked
-- this one takes a long time though, so commenting it, will have no impact on the result
),
wordIndexRatings as( -- reverse the char indexes retrived above so that words occuring earlier have higher weightage
-- and then normalize them to sequential values
select myTableid, Priority, word, ROW_NUMBER() over (partition by myTableid order by wordindex desc) as comparativeWordIndex
from splitWordMatches
)
,
wordIndexSequenceRatings as ( -- need to do this to ensure that if the same set of words from search string is found in two rows,
-- their sequence in the field value is taken into account for higher priority
select w.myTableid, w.word, (w.Priority + w.comparativeWordIndex + coalesce(sequncedPriority ,0)) as Priority
from wordIndexRatings w left join
(
select w1.myTableid, w1.priority, w1.word, w1.comparativeWordIndex, count(w1.myTableid) as sequncedPriority
from wordIndexRatings w1 join wordIndexRatings w2 on w1.myTableId = w2.myTableId and w1.Priority > w2.Priority and w1.comparativeWordIndex>w2.comparativeWordIndex
group by w1.myTableid, w1.priority,w1.word, w1.comparativeWordIndex
)
sequencedPriority on w.myTableId = sequencedPriority.myTableId and w.Priority = sequencedPriority.Priority
),
prioritizedSplitWordMatches as ( -- this calculates the cumulative priority for a field value
select w1.myTableId, sum(w1.Priority) as OverallPriority from wordIndexSequenceRatings w1 join wordIndexSequenceRatings w2 on w1.myTableId = w2.myTableId
where w1.word <> w2.word group by w1.myTableid
),
completeSet as (
select myTableid, priority from plainMatches -- get plain matches which should be highest ranked
union
select myTableid, OverallPriority as priority from prioritizedSplitWordMatches -- get ranked split word matches (which are ordered based on word rank in search string and sequence)
union
select myTableid, Priority as Priority from splitWordMatches -- get one word matches
),
maximizedCompleteSet as( -- set the priority of a field value = maximum priority for that field value
select myTableid, max(priority) as Priority from completeSet group by myTableId
)
select priority, myTable.myTableid , code, Description from maximizedCompleteSet m join myTable on m.myTableId = myTable.myTableId
order by Priority desc, Description -- order by priority desc to get highest rated items on top
--offset 0 rows fetch next 50 rows only -- optional paging

Execute multiple dynamic T-SQL statements and obtain a limited number of unique values while preserving order

I have a SourceTable and a table variable #TQueries containing various T-SQL predicates that target SourceTable.
The expected result is to dynamically generate SELECT statements that return a list of Id's as specified by the predicates in #TQueries. Each dynamically generated SELECT statement also needs to execute in a particular order, and the final set of values needs to be unique and the ordering must be preserved.
Fortunately, there's a limit to how many values need to be retrieved and how many dynamic queries need to be generated. The Id list should contain at most 10 Ids, and we don't expect more than 7 queries.
The following is a sample of this setup, not the actual data/database:
-- Set up some test data, this is quick and dirty just to provide some data to test against
IF NOT EXISTS (SELECT * FROM sys.objects WHERE object_id = OBJECT_ID(N'[dbo].[SourceTable]') AND type in (N'U'))
BEGIN
-- Create a numbers table, sorta
SELECT TOP 20
IDENTITY(INT,1,1) AS Id,
ABS(CHECKSUM(NewId())) % 100 AS [SomeValue]
INTO [SourceTable]
FROM sysobjects a
END
DECLARE #TQueries TABLE (
[Ordinal] INT,
[WherePredicate] NVARCHAR(MAX),
[OrderByPredicate] NVARCHAR(MAX)
);
-- Simulate SELECTs with different order by that get different data due to varying WHERE clauses and ORDER conditions
INSERT INTO #TQueries VALUES ( 1, N'[Id] IN (6,11,13,7,10,3,15)', '[SomeValue] ASC' ) -- Sort Asc
INSERT INTO #TQueries VALUES ( 2, N'[Id] IN (9,15,14,20,17)', '[SomeValue] DESC' ) -- Sort Desc
INSERT INTO #TQueries VALUES ( 3, N'[Id] IN (20,10,1,16,11,19,9,15,17,6,2,3,13)', 'NEWID()' ) -- Sort Random
My main issue has been avoiding the use of a CURSOR or iterating through the rows one by one. The closest I've come to a set operation that meets this criteria is using a table variable to store the results of each query or a massive CTE.
Suggestions and comments are welcome.
Here's a solution that builds a single statement both to run all the queries and to return the results.
It uses a similar approach as in your answer when iterating over the #TQueries table, i.e. it also uses {...} tokens where column values from #TQuery should go, and it puts the values there with nested REPLACE() calls.
Other than that, it heavily depends on ranking functions, and I'm not sure if doesn't really abuse them. You'd need to test this method before deciding if it's better or worse than the one you've got so far.
DECLARE #QueryTemplate nvarchar(max), #FinalSQL nvarchar(max);
SET #QueryTemplate =
N'SELECT
[Id],
QueryRank = {Ordinal},
RowRank = ROW_NUMBER() OVER (ORDER BY {OrderByPredicate})
FROM [dbo].[SourceTable]
WHERE {WherePredicate}
';
SET #FinalSQL =
N'WITH AllData AS (
' +
SUBSTRING(
(
SELECT
'UNION ALL ' +
REPLACE(REPLACE(REPLACE(#QueryTemplate,
'{Ordinal}' , [Ordinal] ),
'{OrderByPredicate}', [OrderByPredicate]),
'{WherePredicate}' , [WherePredicate] )
FROM #TQueries
ORDER BY [Ordinal]
FOR XML PATH (''), TYPE
).value('.', 'nvarchar(max)'),
11, -- starting just after the first 'UNION ALL '
CAST(0x7FFFFFFF AS int) -- max int; no need to specify the exact length
) +
'),
RankedData AS (
SELECT
[Id],
QueryRank,
RowRank,
ValueRank = ROW_NUMBER() OVER (PARTITION BY [Id] ORDER BY QueryRank)
FROM AllData
)SELECT TOP (#top)
[Id]
FROM RankedData
WHERE ValueRank = 1
ORDER BY
QueryRank,
RowRank
';
PRINT #FinalSQL;
EXECUTE sp_executesql #FinalSQL, N'#top int', 10;
Basically, every subquery gets these auxiliary columns:
QueryRank – a constant value (within the subquery's result set) derived from [Ordinal];
RowRank – a ranking assigned to a row based on the [OrderByPredicate].
The result sets are UNIONed and then every entry of every unique value is again ranked (ValueRank) based on the query ranking.
When pulling the final result set, duplicates are suppressed (by the condition ValueRank = 1), and QueryRank and RowRank are used in the ORDER BY clause to preserve the original row order.
I used EXECUTE sp_executesql #query instead of EXECUTE (#query), because the former allows you to add parameters to the query. In particular, I parametrised the number of results to return (the argument of TOP). But you could certainly concatenate that value into the dynamic script directly, just like other things, if you prefer EXECUTE () over EXECUTE sq_executesql.
If you like, you can try this query at SQL Fiddle. (Note: the SQL Fiddle version replaces the #TQueries table variable with the TQueries table.)
This is what I've managed to piece together cobbled from my original response and improved upon by comments from #AndriyM
DECLARE #sql_prefix NVARCHAR(MAX);
SET #sql_prefix =
N'DECLARE #TResults TABLE (
[Ordinal] INT IDENTITY(1,1),
[ContentItemId] INT
);
DECLARE #max INT, #top INT;
SELECT #max = 10;';
DECLARE #sql_insert_template NVARCHAR(MAX), #sql_body NVARCHAR(MAX);
SET #sql_insert_template =
N'SELECT #top = #max - COUNT(*) FROM #TResults;
INSERT INTO #TResults
SELECT TOP (#top) [Id]
FROM [dbo].[SourceTable]
WHERE
{WherePredicate}
AND NOT EXISTS (
SELECT 1
FROM #TResults AS [tr]
WHERE [tr].[ContentItemId] = [SourceTable].[Id]
)
ORDER BY {OrderByPredicate};';
WITH Query ([Ordinal],[SqlCommand]) AS (
SELECT
[Ordinal],
REPLACE(REPLACE(#sql_insert_template, '{WherePredicate}', [WherePredicate]), '{OrderByPredicate}', [OrderByPredicate])
FROM #TQueries
)
SELECT
#sql_body = #sql_prefix + (
SELECT [SqlCommand]
FROM Query
ORDER BY [Ordinal] ASC
FOR XML PATH(''),TYPE).value('.', 'varchar(max)') + CHAR(13)+CHAR(10)
+N' SELECT * FROM #TResults ORDER BY [Ordinal]';
EXEC(#sql_body);
The basic idea is to use a table variable to hold the results of each query. I create a template for the SQL and replace the values in the template based on what is stored in #TQueries.
Once the entire script is completed I run it with EXEC.

Paging, sorting and filtering in a stored procedure (SQL Server)

I was looking at different ways of writing a stored procedure to return a "page" of data. This was for use with the ASP ObjectDataSource, but it could be considered a more general problem.
The requirement is to return a subset of the data based on the usual paging parameters; startPageIndex and maximumRows, but also a sortBy parameter to allow the data to be sorted. Also there are some parameters passed in to filter the data on various conditions.
One common way to do this seems to be something like this:
[Method 1]
;WITH stuff AS (
SELECT
CASE
WHEN #SortBy = 'Name' THEN ROW_NUMBER() OVER (ORDER BY Name)
WHEN #SortBy = 'Name DESC' THEN ROW_NUMBER() OVER (ORDER BY Name DESC)
WHEN #SortBy = ...
ELSE ROW_NUMBER() OVER (ORDER BY whatever)
END AS Row,
.,
.,
.,
FROM Table1
INNER JOIN Table2 ...
LEFT JOIN Table3 ...
WHERE ... (lots of things to check)
)
SELECT *
FROM stuff
WHERE (Row > #startRowIndex)
AND (Row <= #startRowIndex + #maximumRows OR #maximumRows <= 0)
ORDER BY Row
One problem with this is that it doesn't give the total count and generally we need another stored procedure for that. This second stored procedure has to replicate the parameter list and the complex WHERE clause. Not nice.
One solution is to append an extra column to the final select list, (SELECT COUNT(*) FROM stuff) AS TotalRows. This gives us the total but repeats it for every row in the result set, which is not ideal.
[Method 2]
An interesting alternative is given here (https://web.archive.org/web/20211020111700/https://www.4guysfromrolla.com/articles/032206-1.aspx) using dynamic SQL. He reckons that the performance is better because the CASE statement in the first solution drags things down. Fair enough, and this solution makes it easy to get the totalRows and slap it into an output parameter. But I hate coding dynamic SQL. All that 'bit of SQL ' + STR(#parm1) +' bit more SQL' gubbins.
[Method 3]
The only way I can find to get what I want, without repeating code which would have to be synchronized, and keeping things reasonably readable is to go back to the "old way" of using a table variable:
DECLARE #stuff TABLE (Row INT, ...)
INSERT INTO #stuff
SELECT
CASE
WHEN #SortBy = 'Name' THEN ROW_NUMBER() OVER (ORDER BY Name)
WHEN #SortBy = 'Name DESC' THEN ROW_NUMBER() OVER (ORDER BY Name DESC)
WHEN #SortBy = ...
ELSE ROW_NUMBER() OVER (ORDER BY whatever)
END AS Row,
.,
.,
.,
FROM Table1
INNER JOIN Table2 ...
LEFT JOIN Table3 ...
WHERE ... (lots of things to check)
SELECT *
FROM stuff
WHERE (Row > #startRowIndex)
AND (Row <= #startRowIndex + #maximumRows OR #maximumRows <= 0)
ORDER BY Row
(Or a similar method using an IDENTITY column on the table variable).
Here I can just add a SELECT COUNT on the table variable to get the totalRows and put it into an output parameter.
I did some tests and with a fairly simple version of the query (no sortBy and no filter), method 1 seems to come up on top (almost twice as quick as the other 2). Then I decided to test probably I needed the complexity and I needed the SQL to be in stored procedures. With this I get method 1 taking nearly twice as long as the other 2 methods. Which seems strange.
Is there any good reason why I shouldn't spurn CTEs and stick with method 3?
UPDATE - 15 March 2012
I tried adapting Method 1 to dump the page from the CTE into a temporary table so that I could extract the TotalRows and then select just the relevant columns for the resultset. This seemed to add significantly to the time (more than I expected). I should add that I'm running this on a laptop with SQL Server Express 2008 (all that I have available) but still the comparison should be valid.
I looked again at the dynamic SQL method. It turns out I wasn't really doing it properly (just concatenating strings together). I set it up as in the documentation for sp_executesql (with a parameter description string and parameter list) and it's much more readable. Also this method runs fastest in my environment. Why that should be still baffles me, but I guess the answer is hinted at in Hogan's comment.
I would most likely split the #SortBy argument into two, #SortColumn and #SortDirection, and use them like this:
…
ROW_NUMBER() OVER (
ORDER BY CASE #SortColumn
WHEN 'Name' THEN Name
WHEN 'OtherName' THEN OtherName
…
END *
CASE #SortDirection
WHEN 'DESC' THEN -1
ELSE 1
END
) AS Row
…
And this is how the TotalRows column could be defined (in the main select):
…
COUNT(*) OVER () AS TotalRows
…
I would definitely want to do a combination of a temp table and NTILE for this sort of approach.
The temp table will allow you to do your complicated series of conditions just once. Because you're only storing the pieces you care about, it also means that when you start doing selects against it further in the procedure, it should have a smaller overall memory usage than if you ran the condition multiple times.
I like NTILE() for this better than ROW_NUMBER() because it's doing the work you're trying to accomplish for you, rather than having additional where conditions to worry about.
The example below is one based off a similar query I'm using as part of a research query; I have an ID I can use that I know will be unique in the results. Using an ID that was an identity column would also be appropriate here, though.
--DECLARES here would be stored procedure parameters
declare #pagesize int, #sortby varchar(25), #page int = 1;
--Create temp with all relevant columns; ID here could be an identity PK to help with paging query below
create table #temp (id int not null primary key clustered, status varchar(50), lastname varchar(100), startdate datetime);
--Insert into #temp based off of your complex conditions, but with no attempt at paging
insert into #temp
(id, status, lastname, startdate)
select id, status, lastname, startdate
from Table1 ...etc.
where ...complicated conditions
SET #pagesize = 50;
SET #page = 5;--OR CAST(#startRowIndex/#pagesize as int)+1
SET #sortby = 'name';
--Only use the id and count to use NTILE
;with paging(id, pagenum, totalrows) as
(
select id,
NTILE((SELECT COUNT(*) cnt FROM #temp)/#pagesize) OVER(ORDER BY CASE WHEN #sortby = 'NAME' THEN lastname ELSE convert(varchar(10), startdate, 112) END),
cnt
FROM #temp
cross apply (SELECT COUNT(*) cnt FROM #temp) total
)
--Use the id to join back to main select
SELECT *
FROM paging
JOIN #temp ON paging.id = #temp.id
WHERE paging.pagenum = #page
--Don't need the drop in the procedure, included here for rerunnability
drop table #temp;
I generally prefer temp tables over table variables in this scenario, largely so that there are definite statistics on the result set you have. (Search for temp table vs table variable and you'll find plenty of examples as to why)
Dynamic SQL would be most useful for handling the sorting method. Using my example, you could do the main query in dynamic SQL and only pull the sort method you want to pull into the OVER().
The example above also does the total in each row of the return set, which as you mentioned was not ideal. You could, instead, have a #totalrows output variable in your procedure and pull it as well as the result set. That would save you the CROSS APPLY that I'm doing above in the paging CTE.
I would create one procedure to stage, sort, and paginate (using NTILE()) a staging table; and a second procedure to retrieve by page. This way you don't have to run the entire main query for each page.
This example queries AdventureWorks.HumanResources.Employee:
--------------------------------------------------------------------------
create procedure dbo.EmployeesByMartialStatus
#MaritalStatus nchar(1)
, #sort varchar(20)
as
-- Init staging table
if exists(
select 1 from sys.objects o
inner join sys.schemas s on s.schema_id=o.schema_id
and s.name='Staging'
and o.name='EmployeesByMartialStatus'
where type='U'
)
drop table Staging.EmployeesByMartialStatus;
-- Populate staging table with sort value
with s as (
select *
, sr=ROW_NUMBER()over(order by case #sort
when 'NationalIDNumber' then NationalIDNumber
when 'ManagerID' then ManagerID
-- plus any other sort conditions
else EmployeeID end)
from AdventureWorks.HumanResources.Employee
where MaritalStatus=#MaritalStatus
)
select *
into #temp
from s;
-- And now pages
declare #RowCount int; select #rowCount=COUNT(*) from #temp;
declare #PageCount int=ceiling(#rowCount/20); --assuming 20 lines/page
select *
, Page=NTILE(#PageCount)over(order by sr)
into Staging.EmployeesByMartialStatus
from #temp;
go
--------------------------------------------------------------------------
-- procedure to retrieve selected pages
create procedure EmployeesByMartialStatus_GetPage
#page int
as
declare #MaxPage int;
select #MaxPage=MAX(Page) from Staging.EmployeesByMartialStatus;
set #page=case when #page not between 1 and #MaxPage then 1 else #page end;
select EmployeeID,NationalIDNumber,ContactID,LoginID,ManagerID
, Title,BirthDate,MaritalStatus,Gender,HireDate,SalariedFlag,VacationHours,SickLeaveHours
, CurrentFlag,rowguid,ModifiedDate
from Staging.EmployeesByMartialStatus
where Page=#page
GO
--------------------------------------------------------------------------
-- Usage
-- Load staging
exec dbo.EmployeesByMartialStatus 'M','NationalIDNumber';
-- Get pages 1 through n
exec dbo.EmployeesByMartialStatus_GetPage 1;
exec dbo.EmployeesByMartialStatus_GetPage 2;
-- ...etc (this would actually be a foreach loop, but that detail is omitted for brevity)
GO
I use this method of using EXEC():
-- SP parameters:
-- #query: Your query as an input parameter
-- #maximumRows: As number of rows per page
-- #startPageIndex: As number of page to filter
-- #sortBy: As a field name or field names with supporting DESC keyword
DECLARE #query nvarchar(max) = 'SELECT * FROM sys.Objects',
#maximumRows int = 8,
#startPageIndex int = 3,
#sortBy as nvarchar(100) = 'name Desc'
SET #query = ';WITH CTE AS (' + #query + ')' +
'SELECT *, (dt.pagingRowNo - 1) / ' + CAST(#maximumRows as nvarchar(10)) + ' + 1 As pagingPageNo' +
', pagingCountRow / ' + CAST(#maximumRows as nvarchar(10)) + ' As pagingCountPage ' +
', (dt.pagingRowNo - 1) % ' + CAST(#maximumRows as nvarchar(10)) + ' + 1 As pagingRowInPage ' +
'FROM ( SELECT *, ROW_NUMBER() OVER (ORDER BY ' + #sortBy + ') As pagingRowNo, COUNT(*) OVER () AS pagingCountRow ' +
'FROM CTE) dt ' +
'WHERE (dt.pagingRowNo - 1) / ' + CAST(#maximumRows as nvarchar(10)) + ' + 1 = ' + CAST(#startPageIndex as nvarchar(10))
EXEC(#query)
At result-set after query result columns:
Note:
I add some extra columns that you can remove them:
pagingRowNo : The row number
pagingCountRow : The total number of rows
pagingPageNo : The current page number
pagingCountPage : The total number of pages
pagingRowInPage : The row number that started with 1 in this page

Resources