Related
I have 2 strings and one integer:
#categoryID int = 163,
#Ids nvarchar(2000) = '1,2,3',
#Names nvarchar(2000) = 'Bob,Joe,Alex'
I need to select 3 columns 3 rows; The most accomplished is 3 rows 2 columns:
select #categoryID,items from FN_SplitStr(#Ids,',')
resulting:
163,1
163,2
163,3
But I can't figure out how to split both strings.
I tried many ways like:
select #categoryID,items from FN_SplitStr((#Ids,#Names),',')
select #categoryID,items from FN_SplitStr(#Ids,','),items from FN_SplitStr(#Names,',')
EXPECTED OUTPUT:
163,1,Bob
163,2,Joe
163,3,Alex
NOTE1: I looked over tens of questions the most similar is:
How to split string and insert values into table in SQL Server AND SQL Server : split multiple strings into one row each but this question is different.
NOTE2: FN_SplitStr is a function for spliting strings in SQL. And I'm trying to create a stored procedure.
Based on your expected output, you have to use cross apply twice and then create some sort of ranking to make sure that you are getting the right value. As IDs and Names don't seem to have any relationship cross apply will create multiple rows (when you split the string to Names and ID)
There might be better way but this also gives your expected output. You can change this string split to your local function.
1st Dense rank is to make sure that we get three unique names and 2nd dense rank is the rank within the name based on order by with ID and outside of the sub query you have to do some comparison to get only 3 rows.
Declare #categoryID int = 163,
#Ids nvarchar(2000) = '1,2,3',
#Names nvarchar(2000) = 'Bob,Joe,Alex'
select ConcatenatedValue, CategoryID, IDs, Names from (
select concat(#categoryID,',',a.value,',',b.value) ConcatenatedValue, #categoryID CategoryID,
A.value as IDs, b.value as Names , DENSE_RANK() over (order by b.value) as Rn,
DENSE_RANK() over (partition by b.value order by a.value) as Ranked
from string_split(#IDs,',') a
cross apply string_split(#names,',') B ) t
where Rn - Ranked = 0
Output:
Inside your stored procedure do a string split of #Ids and insert into #temp1 table with an identity(1,1) column rowed. You will get:
163,1,1
163,2,2
163,3,3
Then do the second string split of #Names and insert into #temp2 table with an identity(1,1) column rowed. You will get:
Bob,1
Joe,2
Alex,3
You can then do an inner join with #temp1 and #temp2 on #temp1.rowid = #temp2.rowid and get:
163,1,Bob
163,2,Joe
163,3,Alex
I hope this solves your problem.
You can do this with a recursive CTE:
with cte as (
select #categoryId as categoryId,
convert(varchar(max), left(#ids, charindex(',', #ids + ',') - 1)) as id,
convert(varchar(max), left(#names, charindex(',', #names + ',') - 1)) as name,
convert(varchar(max), stuff(#ids, 1, charindex(',', #ids + ','), '')) as rest_ids,
convert(varchar(max), stuff(#names, 1, charindex(',', #names + ','), '')) as rest_names
union all
select categoryId,
convert(varchar(max), left(rest_ids, charindex(',', rest_ids + ',') - 1)) as id,
convert(varchar(max), left(rest_names, charindex(',', rest_names + ',') - 1)) as name,
convert(varchar(max), stuff(rest_ids, 1, charindex(',', rest_ids + ','), '')) as rest_ids,
convert(varchar(max), stuff(rest_names, 1, charindex(',', rest_names + ','), '')) as rest_names
from cte
where rest_ids <> ''
)
select categoryid, id, name
from cte;
Here is a db<>fiddle.
You need to split CSV value with record number. For that you need to use ROW_NUMBER() function to generate record wise unique ID as column like "RID", while you split CSV columns in row.
You can use table value split function or XML as used below.
Please check this let us know your solution is found or not.
DECLARE
#categoryID int = 163,
#Ids nvarchar(2000) = '1,2,3',
#Names nvarchar(2000) = 'Bob,Joe,Alex'
SELECT
#categoryID AS categoryID,
q.Id,
w.Names
FROM
(
SELECT
ROW_NUMBER() OVER (ORDER BY f.value('.','VARCHAR(10)')) AS RID,
f.value('.','VARCHAR(10)') AS Id
FROM
(
SELECT
CAST('<a>' + REPLACE(#Ids,',','</a><a>') + '</a>' AS XML) AS idXML
) x
CROSS APPLY x.idXML.nodes('a') AS e(f)
) q
INNER JOIN
(
SELECT
ROW_NUMBER() OVER (ORDER BY h.value('.','VARCHAR(10)')) AS RID,
h.value('.','VARCHAR(10)') AS Names
FROM
(
SELECT
CAST('<a>' + REPLACE(#Names,',','</a><a>') + '</a>' AS XML) AS namesXML
) y
CROSS APPLY y.namesXML.nodes('a') AS g(h)
) w ON w.RID = q.RID
Actually I m begineer to SQL XML path as so making me professional, Got a scenario...
I have a CTE Function That results as
Data Chars NumberOfOccurance
12 1 1 appears (1 ) times
12 2 2 appears (1 ) times
xx x x appears (2 ) times
and CTE function is :
;with cte as
(
select Data , SUBSTRING(Data,1,1) as Chars,1 as startpos from #t
union all
select Data, SUBSTRING(Data, startpos+1,1) as char,startpos+1 from cte where startpos+1<=LEN(data)
)
select Data,Chars,Cast(Chars as varchar(1)) + ' appears (' + cast(COUNT(*) as varchar(5))+ ' ) times' as 'NumberOfOccurance' from cte
group by data, chars
Actually I just want to make my answer into this :
data Number_of_occurances
12 1 appears (1) times 2 appears (1) times
xx x appears (2) times
I have tries this :
; With Ctea as
(
select Data , SUBSTRING(Data,1,1) as Chars,1 as startpos from #t
union all
select Data, SUBSTRING(Data, startpos+1,1) as char,startpos+1 from ctea where startpos+1<=LEN(data)
)
select Data,Chars,REPLACE((SELECT (Cast(Chars as varchar(1)) + ' appears (' + cast(COUNT(*) as varchar(5))+ ' ) times') AS [data()] FROM Ctea t2 WHERE t2.Data = t1.data FOR XML PATH('')), ' ', ' ;') As Number_of_occurances from ctea as t1
group by t1.data, t1.Chars
It says :
Column 'Ctea.Chars' is invalid in the select list because it is not contained in either an aggregate function or the GROUP BY clause.
when I use temp table and getting my exact answer , but cant do it CTE
Can anyone make my result ?
The problem has nothing to do with your CTE. It actually lies in the following subquery:
SELECT (Cast(Chars as varchar(1)) +
' appears (' +
cast(COUNT(*) as varchar(5)) +
' ) times') AS [data()]
FROM Ctea t2
WHERE t2.Data = t1.data
FOR XML PATH('')
Note how you are using the aggregate COUNT(*) here as well as the column Chars. You need to group by at least Chars here:
SELECT (Cast(Chars as varchar(1)) +
' appears (' +
cast(COUNT(*) as varchar(5)) +
' ) times') AS [data()]
FROM Ctea t2
WHERE t2.Data = t1.data
GROUP BY t2.Chars
FOR XML PATH('')
Furthermore, you do not want to select or group by t1.Chars in the outer query because it will result in one row per value of Chars:
data chars Number_of_occurances
12 1 1 appears (1) times 2 appears (1) times
12 2 1 appears (1) times 2 appears (1) times
xx x x appears (2) times
Finally you should most likely be using the STUFF function, and not REPLACE, as you are trying to create a space-delimited list ("1 appears (1) times 2 appears (1) times"), not replace all space characters with a space and a semicolon ("1 ;appears ;(1) ;times ;2 ;appears ;(1) ;times").
Therefore your final query should be:
; With Ctea as
(
select Data , SUBSTRING(Data,1,1) as Chars,1 as startpos from #t
union all
select Data, SUBSTRING(Data, startpos+1,1) as char,startpos+1 from ctea
where startpos+1<=LEN(data)
)
select Data,
STUFF((SELECT cast(' ' as varchar(max)) + (Cast(Chars as varchar(1)) + ' appears (' + cast(COUNT(*) as varchar(5))+ ' ) times') AS [data()]
FROM Ctea t2
WHERE t2.Data = t1.data
GROUP BY t2.Chars
FOR XML PATH('')), 1, 1, '') As Number_of_occurances
from ctea as t1
group by t1.data
You can use FOR XML PATH like this for concatenation
;with cte as
(
select Data , SUBSTRING(Data,1,1) as Chars,1 as startpos from #t
union all
select Data, SUBSTRING(Data, startpos+1,1) as char,startpos+1 from cte where startpos+1<=LEN(data)
), CTE2 AS
(
select Data,Chars,Cast(Chars as varchar(1)) + ' appears (' + cast(COUNT(*) as varchar(5))+ ' ) times' as 'NumberOfOccurance' from cte
group by data, chars
)
SELECT Data,(SELECT NumberOfOccurance + ' ' FROM CTE2 c2 WHERE c2.Data = C1.Data FOR XML PATH(''),type).value('.','VARCHAR(MAX)') as Number_of_occurances
FROM CTE2 C1
GROUP BY Data
I have this weird scenario (at least it is for me) where I have a table (actually a result set, but I want to make it simpler) that looks like the following:
ID | Actions
------------------
1 | 10,12,15
2 | 11,12,13
3 | 15
4 | 14,15,16,17
And I want to count the different actions in the all the table. In this case, I want the result to be 8 (just counting 10, 11, ...., 17; and ignoring the repeated values).
In case it matters, I am using MS SQL 2008.
If it makes it any easier, the Actions were previously on XML that looks like
<root>
<actions>10,12,15</actions>
</root>
I doubt it makes it easier, but somebody might comeback with an xml function that I am not aware and just makes everything easier.
Let me know if there's something else I should say.
Using approach similar to http://codecorner.galanter.net/2012/04/25/t-sql-string-deaggregate-split-ungroup-in-sql-server/:
First you need a function that would split string, there're many examples on SO, here's one of them:
CREATE FUNCTION dbo.Split (#sep char(1), #s varchar(512))
RETURNS table
AS
RETURN (
WITH Pieces(pn, start, stop) AS (
SELECT 1, 1, CHARINDEX(#sep, #s)
UNION ALL
SELECT pn + 1, stop + 1, CHARINDEX(#sep, #s, stop + 1)
FROM Pieces
WHERE stop > 0
)
SELECT pn,
SUBSTRING(#s, start, CASE WHEN stop > 0 THEN stop-start ELSE 512 END) AS s
FROM Pieces
)
Using this you can run a simple query:
SELECT COUNT(DISTINCT S) FROM MyTable CROSS APPLY dbo.Split(',', Actions)
Here is the demo: http://sqlfiddle.com/#!3/5e706/3/0
SQL Fiddle
MS SQL Server 2008 Schema Setup:
CREATE TABLE Table1
([ID] int, [Actions] varchar(11))
;
INSERT INTO Table1
([ID], [Actions])
VALUES
(1, '10,12,15'),
(2, '11,12,13'),
(3, '15'),
(4, '14,15,16,17')
;
Query 1:
DECLARE #S varchar(255)
DECLARE #X xml
SET #S = (SELECT Actions + ',' FROM Table1 FOR XML PATH(''))
SELECT #X = CONVERT(xml,'<root><s>' + REPLACE(#S,',','</s><s>') + '</s></root>')
SELECT count(distinct [Value])
FROM (
SELECT [Value] = T.c.value('.','varchar(20)')
FROM #X.nodes('/root/s') T(c)) AS Result
WHERE [Value] > 0
Results:
| COLUMN_0 |
|----------|
| 8 |
EDIT :
I think this is exactly what you are looking for :
SQL Fiddle
MS SQL Server 2008 Schema Setup:
Query 1:
DECLARE #X xml
SELECT #X = CONVERT(xml,replace('
<root>
<actions>10,12,15</actions>
<actions>11,12,13</actions>
<actions>15</actions>
<actions>14,15,16,17</actions>
</root>
',',','</actions><actions>'))
SELECT count(distinct [Value])
FROM (
SELECT [Value] = T.c.value('.','varchar(20)')
FROM #X.nodes('/root/actions') T(c)) AS Result
Results:
| COLUMN_0 |
|----------|
| 8 |
A bit if a mess but here it is Create the function first and then call the lower code.
/* Helper Function */
CREATE FUNCTION dbo.Split (#sep char(1), #s varchar(8000))
RETURNS table
AS
RETURN (
WITH splitter_cte AS (
SELECT CHARINDEX(#sep, #s) as pos, 0 as lastPos
UNION ALL
SELECT CHARINDEX(#sep, #s, pos + 1), pos
FROM splitter_cte
WHERE pos > 0
)
SELECT SUBSTRING(#s, lastPos + 1,
case when pos = 0 then 80000
else pos - lastPos -1 end) as chunk
FROM splitter_cte
)
GO
---------------- End of Function
/* Function Call */
Declare #Actions varchar(1000)
SELECT #Actions = STUFF((SELECT ',' + actions
FROM tblActions
ORDER BY actions
FOR XML PATH('')), 1, 1, '')
SELECT Distinct *
FROM dbo.Split(',', #Actions)
OPTION(MAXRECURSION 0);
If you have a table of Actions with one row per possible action id, you can do this with a join:
select count(distinct a.ActionId)
from t join
Actions a
on ','+t.Actions+',' like '%,'+cast(a.ActionId as varchar(255))+',%';
You could also create a table of numbers (using a CTE) if you know the actions are within some range.
I have a bit of a tricky situation. I have a column that contains a pipe delimited set of numbers in numerous rows in a table. For example:
Courses
-------------------
1|2
1|2|3
1|2|8
10
11
11|12
What I want to achieve is to return rows where the number only appears once in my output.
Ideally, I want to try and carry this out using SQL rather than having to carry out checks at a web application level. Carrying out a DISTINCT does not achieve what I want.
The desired output would be:
Courses
-------------------
1
2
3
8
10
11
12
I would appreciated if anyone can guide me in the right direction.
Thanks.
Please try:
declare #tbl as table(Courses nvarchar(max))
insert into #tbl values
('1|2'),
('1|2|3'),
('1|2|8'),
('10'),
('11'),
('11|12')
select * from #tbl
SELECT
DISTINCT CAST(Split.a.value('.', 'VARCHAR(100)') AS INT) AS CVS
FROM
(
SELECT CAST ('<M>' + REPLACE(Courses, '|', '</M><M>') + '</M>' AS XML) AS CVS
FROM #tbl
) AS A CROSS APPLY CVS.nodes ('/M') AS Split(a)
ORDER BY 1
Try this one -
SET NOCOUNT ON;
DECLARE #temp TABLE
(
string VARCHAR(500)
)
DECLARE #Separator CHAR(1)
SELECT #Separator = '|'
INSERT INTO #temp (string)
VALUES
('1|2'),
('1|2|3'),
('1|2|8'),
('10'),
('11'),
('11|12')
-- 1. XML
SELECT p.value('(./s)[1]', 'VARCHAR(500)')
FROM (
SELECT field = CAST('<r><s>' + REPLACE(t.string, #Separator, '</s></r><r><s>') + '</s></r>' AS XML)
FROM #temp t
) d
CROSS APPLY field.nodes('/r') t(p)
-- 2. CTE
;WITH a AS
(
SELECT
start_pos = 1
, end_pos = CHARINDEX(#Separator, t.string)
, t.string
FROM #temp t
UNION ALL
SELECT
end_pos + 1
, CHARINDEX(#Separator, string, end_pos + 1)
, string
FROM a
WHERE end_pos > 0
)
SELECT d.name
FROM (
SELECT
name = SUBSTRING(
string
, start_pos
, ABS(end_pos - start_pos)
)
FROM a
) d
WHERE d.name != ''
Try this :
create table course (courses varchar(100))
insert into course values('1|2')
insert into course values('1|2|3')
insert into course values('1|2|8')
insert into course values('10')
insert into course values('11')
insert into course values('11|12')
Declare #col varchar(200)
SELECT
#col=(
SELECT DISTINCT c.courses + '|'
FROM course c
FOR XML PATH('')
);
select * from course
;with demo as(
select cast(substring(#col,1,charindex('|',#col,1)-1) AS INT) cou,charindex('|',#col,1) pos
union all
select cast(substring(#col,pos+1,charindex('|',#col,pos+1)-pos-1)AS INT) cou,charindex('|',#col,pos+1) pos
from demo where pos<LEN(#col))
select distinct cou from demo
Could not manage without recursion :( Something like this could do the trich?
WITH splitNum(num, r)
AS
(
SELECT
SUBSTRING(<field>,1, CHARINDEX('|', <field>)-1) num,
SUBSTRING(<field>,CHARINDEX('|', <field>)+1, len(<field>)) r
FROM <yourtable> as a
UNION ALL
SELECT
SUBSTRING(r,1, CHARINDEX('|', r)-1) num,
SUBSTRING(r,CHARINDEX('|', r)+1, len(r)) r
FROM <yourtable> b
WHERE CHARINDEX('|', r) > 0
inner join splitNum as c on <whatevertheprimarykeyis>
)
SELECT distinct num FROM splitNum
Didn't make it run, but it should do the trick, just replace the and with the correct info
One way would be to use a recursive CTE:
with cte as
(select cast(case charindex('|',courses) when 0 then courses
else left(courses,charindex('|',courses)-1) end as int) course,
case charindex('|',courses) when 0 then ''
else right(courses,len(courses)-charindex('|',courses)) end courses
from courses
union all
select cast(case charindex('|',courses) when 0 then courses
else left(courses,charindex('|',courses)-1) end as int) course,
case charindex('|',courses) when 0 then ''
else right(courses,len(courses)-charindex('|',courses)) end courses
from cte
where len(courses)>0)
select distinct course from cte
SQLFiddle here.
How can I get each value of a column that has a comma separator in her value ?
Example:
ID ColumnUnified
1 12,34,56,78
2 80,99,70,56
What I want is a query to get the number without comma. If possible, in collumns.
12 34 56 78
This will work for any number of values http://beyondrelational.com/modules/2/blogs/70/posts/10844/splitting-delimited-data-to-columns-set-based-approach.aspx
The solution Madhivanan's link refers to is very creative, but I had a slight problem with it on SQL Server 2012 related to the name of one of the columns (Start). I've modified the code in his answer to use StartPos instead of Start for the column name.
I was not familiar with the system procedure spt_values, but I found a very informative description of the procedure here on SO for those who are interested in exactly how this solution works.
Finally, here's the (slightly) revised code from Madhivana's answer:
CREATE TABLE #test(id int, data varchar(100))
INSERT INTO #test VALUES (1,'This,is,a,test,string')
INSERT INTO #test VALUES (2,'See,if,it,can,be,split,into,many,columns')
DECLARE #pivot varchar(8000)
DECLARE #select varchar(8000)
SELECT #pivot = COALESCE(#pivot + ',', '') + '[col'
+ CAST(number + 1 AS VARCHAR(10)) + ']'
FROM master..spt_values
WHERE type = 'p'
AND number <= ( SELECT MAX(LEN(data) - LEN(REPLACE(data, ',', '')))
FROM #test
)
SELECT #select = '
select p.*
from (
select
id,substring(data, StartPos+2, endPos-StartPos-2) as token,
''col''+cast(row_number() over(partition by id order by StartPos) as varchar(10)) as n
from (
select
id, data, n as StartPos, charindex('','',data,n+2) endPos
from (select number as n from master..spt_values where type=''p'') num
cross join
(
select
id, '','' + data +'','' as data
from
#test
) m
where n < len(data)-1
and substring(data,n+1,1) = '','') as data
) pvt
Pivot ( max(token)for n in (' + #pivot + '))p'
EXEC(#select)
DROP TABLE #test