I have a fairly simple requirement -I have a table with the following (relevant) structure.
with cte as(
select 1 id,'AA,AB,AC,AD' names union all
select 2,'BA,BB' union all
select 3,'CA,CB,CC,CD,CE' union all
select 4,'DA,DB,DC'
)
i would like to create a select statement which will split each "names" column into multiple rows.
For example the first row should produce
1,'AA'
1,'AB'
1,'AC'
1,'AD'
Can we do it using only SQL. This is failry easy to do in Oracle.
You can do it in one query with no custom defined functions if you leverage XML:
WITH cte AS( /*your data*/
SELECT 1 id,'AA,AB,AC,AD' names UNION ALL
SELECT 2,'BA,BB' UNION ALL
SELECT 3,'CA,CB,CC,CD,CE' UNION ALL
SELECT 4,'DA,DB,DC'
)
, xmlData AS ( /*make into xml*/
SELECT id, cast('<root><x>'+replace(names,',','</x><x>')+'</x></root>' as xml) AS theXML
FROM cte
)
SELECT id, x.value('.','varchar(100)') /*split up*/
FROM xmlData
CROSS APPLY xmlData.theXML.nodes('//x') AS func(x)
You can create a split function that returns a table, then select from that table.
/***************************************************************************
**
** Function: split
** In: #ipRowData - The delimited list of items to split.
** In: #ipSplitOn - The delimiter which separates the items in #rowData.
** Returns: A table object containing the split items. The table object
** will have an ID and Data column, where ID is the number of the item
** in the original list and Data is the value of the item.
**
** Description:
** Splits a delimited set of items and returns them
** as a table object.
***************************************************************************/
CREATE FUNCTION [dbo].[split]
(
#ipRowData NVARCHAR(4000),
#ipSplitOn NVARCHAR(5)
)
RETURNS #rtnValue table
(
ID INT identity(1,1),
Data NVARCHAR(100)
)
AS
BEGIN
DECLARE
#cnt INT
Set #cnt = 1
WHILE (Charindex(#ipSplitOn,#ipRowData)>0)
BEGIN
INSERT INTO #rtnValue
( data )
SELECT Data = ltrim(rtrim(Substring(#ipRowData,1,Charindex(#ipSplitOn,#ipRowData)-1)))
SET #ipRowData = Substring(#ipRowData,Charindex(#ipSplitOn,#ipRowData)+1,len(#ipRowData))
SET #cnt = #cnt + 1
END
INSERT INTO #rtnValue (data)
SELECT DATA = ltrim(rtrim(#ipRowData))
RETURN
END
GO
Sample Usage:
select 1,data from [dbo].split('AA,AB,AC,AD', ',');
Output:
(No column name) data
1 AA
1 AB
1 AC
1 AD
Related
I am trying to create a stored procedure in SQL Server which will be used from C# Entity Framework.
My main focus is take a input of long string text then split that data by characters and return list of values of matching data.
In detail:
- is separator between name of data type and its value
: is separator between Type and ASIN
, is separator between two different value
I want to get List of data filtered by ASIN and Type from this stored procedure. I am getting full text string in the #DataString variable but I don't know how I can split my text and run SELECT to return all of data.
Any idea to do it? Ask any question you may have.
Example of long text string:
Type-1:ASIN-NsQf8,ASIN-YhQfu,ASIN-dpQf9,ASIN-rsWf3
The unfinished SQL code:
CREATE PROCEDURE dbo.lk_GetMatchingDataOfThirdparty
#DataString VARCHAR(MAX)
AS
BEGIN
SET NOCOUNT ON;
SELECT *
FROM ThirdPartyData
WHERE ASIN = '#value_get_from_string'
AND Type = '#value_get_from_string'
END
Use a split function:
CREATE FUNCTION [dbo].[split](
#delimited NVARCHAR(MAX),
#delimiter NVARCHAR(100)
) RETURNS #t TABLE (id INT IDENTITY(1,1), val NVARCHAR(MAX))
AS
BEGIN
DECLARE #xml XML
SET #xml = N'<t>' + REPLACE(#delimited,#delimiter,'</t><t>') + '</t>'
INSERT INTO #t(val)
SELECT r.value('.','nvarchar(MAX)') as item
FROM #xml.nodes('/t') as records(r)
RETURN
END
GO
DECLARE #DataString VARCHAR(MAX);
SET #DataString ='Type-1:ASIN-NsQf8,ASIN-YhQfu,ASIN-dpQf9,ASIN-rsWf3'
;WITH cte as (
SELECT a.id as [1], b.id as [2], c.id as [3], c.val
FROM (
SELECT * FROM dbo.split(#DataString, ':')
) a
CROSS APPLY dbo.split(a.val,',') b
CROSS APPLY dbo.split(b.val,'-') c
),
typecte as (
select b.val as [TypeValue]
from cte a
inner join cte b
ON a.[1] = b.[1]
AND a.[2] = b.[2]
AND a.[3]+1 = b.[3] -- Next value
WHERE a.val='Type'
),
asincte as (
select b.val as [ASINValue]
from cte a
inner join cte b
ON a.[1] = b.[1]
AND a.[2] = b.[2]
AND a.[3]+1 = b.[3] -- Next value
WHERE a.val='ASIN'
)
SELECT *
FROM ThirdPartyData
WHERE [ASIN] IN (SELECT [ASINValue] FROM asincte)
AND [Type] IN (SELECT [TypeValue] FROM typecte)
An oft overlooked way of doing coding that isn't really relational database related is a SQL CLR. String manipulation is a good example of something a SQL CLR could handle much better than a SQL script. What you could do in the above example is call a SQL CLR function that does the string manipulation on the long string text to return the values you need then plug those variables into your SELECT statement.
I have the following SQL which queries a single table, single row, and returns the results as a comma separate string e.g.
Forms
1, 10, 4
SQL :
DECLARE #tmp varchar(250)
SET #tmp = ''
SELECT #tmp = #tmp + Form_Number + ', '
FROM Facility_EI_Forms_Required
WHERE Facility_ID = 11 AND EI_Year=2012 -- single Facility, single year
SELECT SUBSTRING(#tmp, 1, LEN(#tmp) - 1) AS Forms
The Facility_EI_Forms_Required table has three records for Facility_ID = 11
Facility_ID EI_Year Form_Number
11 2012 1
11 2012 10
11 2012 4
Form_number is a varchar field.
And I have a Facility table with Facility_ID and Facility_Name++.
How do I create a query to query all Facilites for a given year and produce the CSV output field?
I have this so far:
DECLARE #tmp varchar(250)
SET #tmp = ''
SELECT TOP 100 A.Facility_ID, A.Facility_Name,
(
SELECT #tmp = #tmp + B.Form_Number + ', '
FROM B
WHERE B.Facility_ID = A.Facility_ID
AND B.EI_Year=2012
)
FROM Facility A, Facility_EI_Forms_Required B
But it gets syntax errors on using #tmp
My guess is this is too complex a task for a query and a stored procedure may be need, but I have little knowledge of SPs. Can this be done with a nested query?
I tried a Scalar Value Function
ALTER FUNCTION [dbo].[sp_func_EI_Form_List]
(
-- Add the parameters for the function here
#p1 int,
#pYr int
)
RETURNS varchar
AS
BEGIN
-- Declare the return variable here
DECLARE #Result varchar
-- Add the T-SQL statements to compute the return value here
DECLARE #tmp varchar(250)
SET #tmp = ''
SELECT #tmp = #tmp + Form_Number + ', '
FROM OIS..Facility_EI_Forms_Required
WHERE Facility_ID = #p1 AND EI_Year = #pYr -- single Facility, single year
SELECT #Result = #tmp -- SUBSTRING(#tmp, 1, LEN(#tmp) - 1)-- #p1
-- Return the result of the function
RETURN #Result
END
The call
select Facility_ID, Facility.Facility_Name,
dbo.sp_func_EI_Form_List(Facility_ID,2012)
from facility where Facility_ID=11
returns
Facility_ID Facility_Name Form_List
11 Hanson Aggregates 1
so it is only returning the first record instead of all three. What am I doing wrong?
Try the following approach, which is an analogy to SO answer Concatenate many rows into a single text string. I hope it is correct, as I cannot try it out without having the schema and some demo data (maybe you can add schema and data to your question):
Select distinct A.Facility_ID, A.Facility_Name,
substring(
(
Select ',' + B.Form_Number AS [text()]
From Facility_EI_Forms_Required B
Where B.Facility_ID = A.Facility_ID
AND B.EI_Year=2012
ORDER BY B.Facility_ID
For XML PATH ('')
), 2, 1000) [Form_List]
From Facility A
I have a table with 3 columns.
one of them is [Code]. I have many records on this table.
I want to select records that their [Code] are numbers close to 10 regularly
for example if select records that has [Code]=9 then select records that has [Code] = 8 etc...
This is what I implement based on your though.
If you wish near record or record-id, not value, then you can change only condition a.data to a.rid.
declare #t table (data int)
insert into #t values(1), (2),(3),(4),(5),(6),(7),(8),(9),(10),(11),(12),(50),(51),(52)
declare #value int = 11 , #getDatToValue int = 2
select * from
(
select * , ROW_NUMBER( ) over(order by data) rid
from #t
)
a
where
a.data between (#value - #getDatToValue) and (#value + #getDatToValue)
I want to split each name for individual columns
create table split_test(value integer,Allnames varchar(40))
insert into split_test values(1,'Vinoth,Kumar,Raja,Manoj,Jamal,Bala');
select * from split_test;
Value Allnames
-------------------
1 Vinoth,Kumar,Raja,Manoj,Jamal,Bala
Expected output
values N1 N2 N3 N4 N5 N6 N7.......N20
1 Vinoth Kumar Raja Manoj Jamal Bala
using this example you can get an idea.
declare #str varchar(max)
set #str = 'Hello world'
declare #separator varchar(max)
set #separator = ' '
declare #Splited table(id int identity(1,1), item varchar(max))
set #str = REPLACE(#str,#separator,'''),(''')
set #str = 'select * from (values('''+#str+''')) as V(A)'
insert into #Splited
exec(#str)
select * from #Splited
Here is an sql statement using recursive CTE to split names into rows, then pivot rows into columns.
SqlFiddle
with names as
(select
value,
1 as name_id,
substring(Allnames,1,charindex(',',Allnames+',', 0)-1) as name,
substring(Allnames,charindex(',',Allnames, 0)+1, 40) as left_names
from split_test
union all
select
value,
name_id +1,
case when charindex(',',left_names, 0)> 0 then
substring(left_names,1,charindex(',',left_names, 0)-1)
else left_names end as name,
case when charindex(',',left_names, 0)> 0 then
substring(left_names,charindex(',',left_names, 0)+1, 40)
else '' end as left_names
from names
where ltrim(left_names)<>'')
select value,
[1],[2],[3],[4],[5],[6],[7],[8],[9]
from (select value,name_id,name from names) as t1
PIVOT (MAX(name) FOR name_id IN ( [1],[2],[3],[4],[5],[6],[7],[8],[9] ) ) AS t2
UPDATE
#KM.'s answer might be a better way to split data into rows without recursive CTE table. It should be more efficient than this one. So I follow that example and simplified the part of null value process logic. Here is the result:
Step 1:
Create a table includes all numbers from 1 to a number grater than max length of Allnames column.
CREATE TABLE Numbers( Number int not null primary key);
with n as
(select 1 as num
union all
select num +1
from n
where num<100)
insert into numbers
select num from n;
Step 2:
Join data of split_test table with numbers table, we can get all the parts start from ,.
Then take the first part between 2 , form every row. If there are null values exists, add them with union.
select value ,
ltrim(rtrim(substring(allnames,number+1,charindex(',',substring(allnames,number,40),2)-2))) as name
from
(select value, ','+allnames+',' as allnames
from split_test) as t1
left join numbers
on number<= len(allnames)
where substring(allnames,number,1)=','
and substring(allnames,number,40)<>','
union
select value, Allnames
from split_test
where Allnames is null
Step 3: Pivot names from rows to columns like my first attempt above, omitted here.
SQLFiddle
CREATE FUNCTION [dbo].[udfGetNextEntityID]
()
RETURNS INT
AS
BEGIN
;WITH allIDs AS
(
SELECT entity_id FROM Entity
UNION SELECT entity_id FROM Reserved_Entity
)
RETURN (SELECT (MAX(entity_id) FROM allIDs )
END
GO
SQL isn't my strong point, but I can't work out what I'm doing wrong here. I want the function to return the largest entity_id from a union of 2 tables. Running the script gives the error:
Incorrect syntax near the keyword 'RETURN'.
I looked to see if there was some restriction on using CTEs in functions but couldn't find anything relevant. How do I correct this?
CREATE FUNCTION [dbo].[udfGetNextEntityID]()
RETURNS INT
AS
BEGIN
DECLARE #result INT;
WITH allIDs AS
(
SELECT entity_id FROM Entity
UNION SELECT entity_id FROM Reserved_Entity
)
SELECT #result = MAX(entity_id) FROM allIDs;
RETURN #result;
END
GO
While you can do it, why do you need a CTE here?
RETURN
(
SELECT MAX(entity_id) FROM
(
SELECT entity_id FROM dbo.Entity
UNION ALL
SELECT entity_id FROM dbo.Reserved_Entity
) AS allIDs
);
Also there is no reason to use UNION instead of UNION ALL since this will almost always introduce an expensive distinct sort operation. And please always use the schema prefix when creating / referencing any object.
You can not return the way your are doing from the function.
Make use of a local variable and return the same.
CREATE FUNCTION [dbo].[udfGetNextEntityID]()
RETURNS INT
AS
BEGIN
DECLARE #MaxEntityId INT;
WITH allIDs AS
(
SELECT entity_id FROM Entity
UNION SELECT entity_id FROM Reserved_Entity
)
SELECT #MaxEntityId = MAX(entity_id) FROM allIDs;
RETURN #MaxEntityId ;
END
GO
create function tvfFormatstring (#string varchar(100))
returns #fn_table table
(id int identity(1,1),
item int)
as
begin
insert into #fn_table(item)
declare #result int
set #string = #string+'-'
;with cte (start,number)
as
(
select 1 as start , CHARINDEX('-',#string,1) as number
union all
select number+1 as start , CHARINDEX('-',#string,number+1) as number from cte
where number <= LEN(#string)
)
select #result = SUBSTRING(#string,start,number-start) from cte ;
return #result;
end
select * from tvfFormatstring ('12321-13542-15634')