I have a SQL Query that I need to return rows based on the following possibilities:
ReqType
------
NULL
LTL
TL
To get a better idea of what I want to achieve:
SELECT * FROM MyTable MT
WHERE MT.ReqType = #param1
The parameter that is being passed is #param. Possibilities for #param can be NULL, TL or LTL.
#param and ReqType are nvarchar(3)
NULL doesn't work with =, so you need more explicit logic:
WHERE (MT.ReqType = #param1 OR
(MT.ReqType IS NULL AND #param1 IS NULL)
)
If #param1 (can be) SQL NULL then you could also do:
SELECT * FROM MyTable MT
WHERE isnull(MT.ReqType, '') = isnull(#param1,'')
*That would not work if you are allowing empty string for your MT.ReqType. In that case you could pick another char to replace null values with.
Works the same as Gordon's, a little more compact, note the use of coalesce. I don't think this is Index friendly so you might want to be careful how you use it.
declare #test table
(
ReqFld1 varchar(4),
ReqType varchar(3)
)
insert into #test values ('FLD1', 'LTL')
insert into #test values ('FLD2', 'TL')
insert into #test values ('FLD3', NULL)
declare #srch varchar(3)
select #srch = null
select * from #test
where coalesce(ReqType, '') = (coalesce(#srch, ''))
select #srch = 'LTL'
select * from #test
where coalesce(ReqType, '') = (coalesce(#srch, ''))
select #srch = 'TL'
select * from #test
where coalesce(ReqType, '') = (coalesce(#srch, ''))
Related
I am trying to implement search functionality with list of values in SQL variable, including range. Appreciate any guidance/links pointing to correct approach for this.
Below is the dataset:
CREATE TABLE [dbo].[Books]
(
[ID] [NCHAR](10) NOT NULL,
[AUTHCODE] [NCHAR](10) NULL,
[TITLE] [NCHAR](10) NULL
) ON [PRIMARY]
GO
INSERT [dbo].[Books] ([ID], [AUTHCODE], [TITLE])
VALUES (N'1', N'nk', N'Book1'),
(N'2', N'an', N'Book2'),
(N'3', N'mn', N'Book3'),
(N'4', N'ra', N'Book4'),
(N'5', N'kd', N'Book5'),
(N'6', N'nk', N'Book6'),
(N'7', N'an', N'Book7'),
(N'8', N'ra', N'Book8'),
(N'9', N'kd', N'Book9'),
(N'10', N'mn', N'Book10 ')
GO
Below I am trying to filter using the SQL IN clause but this does not return desired result.
select * from books
declare #List1 varchar(max) = '2,4,6,7,8,9' --simple list
select *
from books
where id in (#List1)
declare #List2 varchar(max) = '2,4-7,9' --list with range
select *
from books
where id in (#List2)
You cannot directly use strings as lists, but you can do use STRING_SPLIT (Transact-SQL) if you really need to pass filtering parameters as strings:
declare #list varchar(max) = '2,4,6-8,9'
declare #filter table (id1 int, id2 int)
insert into #filter (id1,id2)
select
case when b.pos > 0 then left(a.[value], pos - 1) else a.[value] end as id1,
case when b.pos > 0 then right(a.[value], len(a.[value]) - pos) else a.[value] end as id2
from string_split(#list, ',') as a
cross apply (select charindex('-', a.[value]) as pos) as b
select *
from [dbo].[Books] as b
where
exists (select * from #filter as tt where b.id between tt.id1 and tt.id2)
Also it might be an idea to pass your filter as json and OPENJSON (Transact-SQL) so you can make parsing part simplier:
declare #list varchar(max) = '[2,4,[6,8],9]'
select
case when a.[type] = 4 then json_value(a.[value], '$[0]') else a.[value] end,
case when a.[type] = 4 then json_value(a.[value], '$[1]') else a.[value] end
from openjson(#list) as a
All above, of course, only applicable if you have Sql Server 2016 or higher
IN() operator determines whether a specified value matches any value in a subquery or a list. not a string and that what you are doing.
What you are trying to do can be done as
DECLARE #List1 AS TABLE (ID INT);
INSERT INTO #List1
SELECT 2 UNION SELECT 4 UNION SELECT 6 UNION
SELECT 7 UNION SELECT 8 UNION SELECT 9;
SELECT *
FROM Books
WHERE ID IN (SELECT ID FROM #List1);
DECLARE #List2 As TABLE (AFrom INT, ATo INT);
INSERT INTO #List2
SELECT 2, 4 UNION SELECT 7, 9;
SELECT *
FROM Books B CROSS APPLY #List2 L
WHERE B.ID BETWEEN L.AFrom AND L.ATo;
Live Demo
I want to select records from a table in a stored procedure. Given parameters can be empty or a string including some keys separated by comma (1, 2, etc)
I want to manage that when a parameter is an empty string, "WHERE" ignore searching.
I'm using this code:
where (CASE when #PatientID <> 0 then ( dental.ID_Sick in (1,2)) else (1=1) end)
Something like that is working in W3School. I mean:
SELECT * FROM Customers
WHERE (case when 1=1 then (Country IN ('Germany', 'France', 'UK')) else 1=1 end);
What is the problem in my query that does not work? SQLServerManagementStudio is giving error on "IN" statement.
Solution:
The best way to handle such optional parameters is to use dynamic SQL and built the query on the fly. Something like....
CREATE PROCEDURE myProc
#Param1 VARCHAR(100) = NULL
,#Param2 VARCHAR(100) = NULL
,#Param3 VARCHAR(100) = NULL
,#ListParam VARCHAR(100) = NULL
--, etc etc...
AS
BEGIN
SET NOCOUNT ON;
Declare #Sql NVARCHAR(MAX);
SET #Sql = N' SELECT *
FROM TableName
WHERE 1 = 1 '
-- add in where clause only if a value was passed to parameter
+ CASE WHEN #Param1 IS NOT NULL THEN
N' AND SomeColumn = #Param1 ' ELSE N'' END
-- add in where clause a different variable
-- only if a value was passed to different parameter
+ CASE WHEN #Param2 IS NOT NULL THEN
N' AND SomeOtherColumn = #Param3 ' ELSE N'' END
-- List Parameter used with IN clause if a value is passed
+ CASE WHEN #ListParam IS NOT NULL THEN
N' AND SomeOtherColumn IN (
SELECT Split.a.value(''.'', ''VARCHAR(100)'') IDs
FROM (
SELECT Cast (''<X>''
+ Replace(#ListParam, '','', ''</X><X>'')
+ ''</X>'' AS XML) AS Data
) AS t CROSS APPLY Data.nodes (''/X'') AS Split(a) '
ELSE N'' END
Exec sp_executesql #sql
, N' #Param1 VARCHAR(100), #Param2 VARCHAR(100) ,#Param3 VARCHAR(100) ,#ListParam VARCHAR(100)'
, #Param1
, #Param2
,#Param3
, #ListParam
END
Problem with Other approach
There is a major issue with this other approach, you write your where clause something like...
WHERE ( ColumnName = #Parameter OR #Parameter IS NULL)
The Two major issues with this approach
1) you cannot force SQL Server to check evaluate an expression first like if #Parameter IS NULL, Sql Server might decide to evaluate first the expression ColumnName = #Parameterso you will have where clause being evaluated even if the variable value is null.
2) SQL Server does not do Short-Circuiting (Like C#), even if it decides to check the #Parameter IS NULL expression first and even if it evaluates to true, SQL Server still may go ahead and evaluating other expression in OR clause.
Therefore stick to Dynamic Sql for queries like this. and happy days.
SQL Server does not have a Bool datatype, so you can't assign or return the result of a comparison as a Bool as you would in other languages. A comparison can only be used with IF-statements or WHERE-clauses, or in the WHEN-part of a CASE...WHEN but not anywhere else.
Your specific example would become this:
SELECT * FROM Customers
WHERE 1=1 OR Country IN ('Germany', 'France', 'UK')
It would be better readable to rewrite your statement as follows:
WHERE #PatientID = 0
OR dental.ID_Sick in (1,2)
Referring to your actual question, I'd advise to read the linked question as provided by B House.
May be this straight way will work for you
IF (#PatientID <> 0)
BEGIN
SELECT * FROM Customers
WHERE Country IN ('Germany', 'France', 'UK')
END
try this:
WHERE 1=(CASE WHEN #PatientID <>0 AND dental.ID_Sick in (1,2) THEN 1
WHEN #PatientID =0 THEN 1
ELSE 0
END)
I have a table UserPermission which has a number of columns of TINYINT type. e.g Read, Write, Update, Delete, Access etc.
I get three parameters in the stored procedure: #UserId, #ColNames, #ColValues where #ColNames and #ColValues are comma separated values.
How can I insert or update the table row (if already exists) with the passed column names and corresponding values.
I try to write the dynamic query which runs fine for INSERT but I was unable to write the UPDATE query dynamically with each column and its value to be concatenate.
Any response would be appreciated
Thanks in advance.
This is a somewhat dirty way to do what you require. However, if you create the following Stored Procedure:
CREATE FUNCTION [dbo].[stringSplit]
(
#String NVARCHAR(4000),
#Delimiter NCHAR(1)
)
RETURNS TABLE
AS
RETURN
(
WITH Split(stpos,endpos)
AS(
SELECT 0 AS stpos, CHARINDEX(#Delimiter,#String) AS endpos
UNION ALL
SELECT endpos+1, CHARINDEX(#Delimiter,#String,endpos+1)
FROM Split
WHERE endpos > 0
)
SELECT 'Id' = ROW_NUMBER() OVER (ORDER BY (SELECT 1)),
'Data' = SUBSTRING(#String,stpos,COALESCE(NULLIF(endpos,0),LEN(#String)+1)-stpos)
FROM Split
)
You can then use that Procedure to join the data together:
DECLARE #TotalCols INT
DECLARE #TotalVals INT
SET #TotalCols = (
SELECT COUNT(ID) AS Total
FROM dbo.stringSplit('department, teamlead', ',')
);
SET #TotalVals = (
SELECT COUNT(ID) AS Total
FROM dbo.stringSplit('IT, Bob', ',')
);
IF #TotalCols = #TotalVals
BEGIN
IF OBJECT_ID('tempdb..#temptable') IS NOT NULL
DROP TABLE #temptable
CREATE TABLE #temptable (
ColName VARCHAR(MAX) NULL
,ColValue VARCHAR(MAX) NULL
)
INSERT INTO #temptable
SELECT a.DATA
,b.DATA
FROM dbo.stringSplit('department, teamlead', ',') AS a
INNER JOIN dbo.stringSplit('IT, Bob', ',') AS b ON a.Id = b.Id
SELECT *
FROM #temptable;
END
It's not very efficient, but it will bring you the desired results.
You can then use the temp table to update, insert and delete as required.
Instead of having a comma delimited list I would create a separate parameter for each Column and make its default value to NULL and in the code update nothing if its null or insert 0. Something like this....
CREATE PROCEDURE usp_UserPermissions
#UserID INT
,#Update INT = NULL --<-- Make default values NULL
,#Delete INT = NULL
,#Read INT = NULL
,#Write INT = NULL
,#Access INT = NULL
AS
BEGIN
SET NOCOUNT ON;
Declare #t TABLE (UserID INT, [Update] INT,[Read] INT
,[Write] INT,[Delete] INT,[Access] INT)
INSERT INTO #t (Userid, [Update],[Read],[Write],[Delete],[Access])
VALUES (#UserID , #Update , #Read, #Write , #Delete, #Access)
IF EXISTS (SELECT 1 FROM UserPermission WHERE UserID = #UserID)
BEGIN
UPDATE up -- Only update if a value was provided else update to itself
SET up.[Read] = ISNULL(t.[Read] , up.[Read])
,up.[Write] = ISNULL(t.[Write] , up.[Write])
,up.[Update] = ISNULL(t.[Update] , up.[Update])
,up.[Delete] = ISNULL(t.[Delete] , up.[Delete])
,up.[Access] = ISNULL(t.[Access] , up.[Access])
FROM UserPermission up
INNER JOIN #t t ON up.UserID = t.UserID
END
ELSE
BEGIN
-- if already no row exists for that User add a row
-- If no value was passed for a column add 0 as default
INSERT INTO UserPermission (Userid, [Update],[Read],[Write],[Delete],[Access])
SELECT Userid
, ISNULL([Update], 0)
, ISNULL([Read], 0)
, ISNULL([Write], 0)
, ISNULL([Delete], 0)
, ISNULL([Access], 0)
FROM #t
END
END
I have two tables named Retail and Activity and the data is as shown below:
Retail Table
Activity Table
My main concern is about Ok and Fault column of the table Retail, as you can see it contains comma separated value of ActivityId.
What i want is, if the Ok column has ActivityId the corresponding column will have Yes, if the Fault column has ActivityId then it should be marked as No
Note I have only four columns that is fixed, it means i have to check that either four of the columns has its value in Ok or Fault, if yes then only i have to print yes or no, otherwise null.
Desired result should be like :
If the value is in Ok then yes other wise No.
I guessing you want to store 'yes' or 'No' in some column. Below is the query to update that column :
UPDATE RetailTable
SET <Result_Column>=
CASE
WHEN Ok IS NOT NULL THEN 'Yes'
WHEN Fault IS NOT NULL THEN 'No'
END
You can use below code as staring point:
DECLARE #Retail TABLE
(
PhoneAuditID INT,
HandsetQuoteID INT,
Ok VARCHAR(50)
)
INSERT INTO #Retail VALUES (1, 1009228, '4,22,5')
INSERT INTO #Retail VALUES (2, 1009229, '1')
DECLARE #Activity TABLE
(
ID INT,
Activity VARCHAR(50)
)
INSERT INTO #Activity VALUES (1, 'BatteryOK?'), (4, 'PhonePowersUp?'), (22,'SomeOtherQuestion?'), (5,'LCD works OK?')
SELECT R.[PhoneAuditID], R.[HandsetQuoteID], A.[Activity], [Ok] = CASE WHEN A.[ID] IS NOT NULL THEN 'Yes' END
FROM #Retail R
CROSS APPLY dbo.Split(R.Ok, ',') S
LEFT JOIN #Activity A ON S.[items] = A.[ID]
I have used Split function provided here:
separate comma separated values and store in table in sql server
Try following query. i have used pivot to show row as columns. I have also used split function to split id values which you can find easily on net:
CREATE TABLE PhoneAudit
(
PhoneAuditRetailID INT,
HandsetQuoteID INT,
Ok VARCHAR(50),
Fault VARCHAR(50)
)
INSERT INTO PhoneAudit VALUES (1,10090,'1,2','3')
CREATE TABLE ActivityT
(
ID INT,
Activity VARCHAR(100)
)
INSERT INTO ActivityT VALUES (1,'Battery')
INSERT INTO ActivityT VALUES (2,'HasCharger')
INSERT INTO ActivityT VALUES (3,'HasMemoryCard')
INSERT INTO ActivityT VALUES (4,'Test')
DECLARE #SQL AS NVARCHAR(MAX)
DECLARE #ColumnName AS NVARCHAR(MAX)
SELECT #ColumnName= ISNULL(#ColumnName + ',','') + QUOTENAME(Activity) FROM (SELECT DISTINCT Activity FROM ActivityT) AS Activities
SET #SQL = 'SELECT PhoneAuditRetailID, HandsetQuoteID,
' + #ColumnName + '
FROM
(SELECT
t1.PhoneAuditRetailID,
t1.HandsetQuoteID,
TEMPOK.*
FROM
PhoneAudit t1
CROSS APPLY
(
SELECT
Activity,
(CASE WHEN ID IN (SELECT * FROM dbo.SplitIDs(t1.Ok,'',''))
THEN ''YES''
ELSE ''NO''
END) AS VALUE
FROM
ActivityT t2
) AS TEMPOK) AS t3
PIVOT
(
MIN(VALUE)
FOR Activity IN ('+ #ColumnName + ')
) AS PivotTable;'
EXEC sp_executesql #SQL
DROP TABLE PhoneAudit
DROP TABLE ActivityT
There are several ways to do this. If you are looking for a purely declarative approach, you could use a recursive CTE. The following example of this is presented as a generic solution with test data which should be adaptable to your needs:
Declare #Delimiter As Varchar(2)
Set #Delimiter = ','
Declare #Strings As Table
(
String Varchar(50)
)
Insert Into #Strings
Values
('12,345,6,78,9'),
(Null),
(''),
('123')
;With String_Columns As
(
Select
String,
Case
When String Is Null Then ''
When CharIndex(#Delimiter,String,0) = 0 Then ''
When Len(String) = 0 Then ''
Else Left(String,CharIndex(#Delimiter,String,0)-1)
End As String_Column,
Case
When String Is Null Then ''
When CharIndex(#Delimiter,String,0) = 0 Then ''
When Len(String) = 0 Then ''
When Len(Left(String,CharIndex(#Delimiter,String,0)-1)) = 0 Then ''
Else Right(String,Len(String)-Len(Left(String,CharIndex(#Delimiter,String,0)-1))-1)
End As Remainder,
1 As String_Column_Number
From
#Strings
Union All
Select
String,
Case
When CharIndex(#Delimiter,Remainder,0) = 0 Then Remainder
Else Left(Remainder,CharIndex(#Delimiter,Remainder,0)-1)
End As Remainder,
Case
When CharIndex(#Delimiter,Remainder,0) = 0 Then ''
When Len(Left(Remainder,CharIndex(#Delimiter,Remainder,0)-1)) = 0 Then ''
Else Right(Remainder,Len(Remainder)-Len(Left(Remainder,CharIndex(#Delimiter,Remainder,0)-1))-1)
End As Remainder,
String_Column_Number + 1
From
String_Columns
Where
(Remainder Is Not Null And Len(Remainder) > 1)
)
Select
String,
String_Column,
String_Column_Number
From
String_Columns
Consider the following table:
DECLARE #tmp_Example
TABLE (
TestInt int,
TestName varchar(max)
);
INSERT INTO #tmp_Example (TestInt, TestName)
VALUES (22,'Bob'),
(23,'James'),
(24,'Joe');
SELECT * from #tmp_Example
WHERE TestInt = 23;
I am getting into situations where somebody might want to select multiple results for TestInt, and I am receiving it like so:
('23,24')
Without changing the schema of the table, how can I support such a query? I fiddled with splitting it like so:
DECLARE #testEx as varchar(max) = '23,24';
SELECT * from #tmp_Example
WHERE TestInt = CAST(dbo.Split(#testEx,',') AS int);
Assuming dbo.Split is what I think it is, a table-valued function that will return integers 23 and 24 in their own rows:
SELECT t.TestInt, t.TestName
FROM #tmp_Example AS t
INNER JOIN dbo.Split(#testEx, ',') AS s
ON t.TestInt = s.OutputColumnName;
SQLfiddle demo
This approach works also (assuming you handle the case where someone has spaces in their list, e.g. 23, 24):
SELECT TestInt, TestName
FROM #tmp_Example
WHERE ','+REPLACE(#testEx, ' ', '')+','
LIKE '%,'+cast(TestInt as varchar(255))+',%';
SQLfiddle demo