Comparing data between rows within a table - tsql

I have following table structure
table 1
ID SOURCE_ID NAME
1 1 A
2 1 B
3 2 B
4 2 C
5 2 A
i need to pick those names which are common across all SOURCE_ID , hence i expect names A and B as they are present in both the SOURCE_ID 1,2.
The following query gives me the expected output:
SELECT DISTINCT NAME
FROM TABLE1 A, TABLE1 B
WHERE A.NAME = B.NAME AND A.SOURCE_ID != B.SOURCE_ID
Now when the data in table changes to include a new record ID 6
table 1
ID SOURCE_ID NAME
1 1 A
2 1 B
3 2 B
4 2 C
5 2 A
6 3 A
The name that is common in all three SOURCE_ID(1,2,3) IS A.
My query fails to return the correct output as new records are entered.
Please provide me a query that works correctly when new records are inserted.

Have a look at something like
DECLARE #Table TABLE(
SOURCE_ID INT,
NAME VARCHAR(20)
)
INSERT INTO #Table SELECT 1,'A'
INSERT INTO #Table SELECT 1,'B'
INSERT INTO #Table SELECT 2,'B'
INSERT INTO #Table SELECT 2,'C'
INSERT INTO #Table SELECT 2,'A'
--INSERT INTO #Table SELECT 3,'A'
;WITH DistinctCount AS (
SELECT NAME,
COUNT(DISTINCT SOURCE_ID) Cnt
FROM #Table
GROUP BY NAME
)
SELECT *
FROM DistinctCount
WHERE Cnt = (SELECT COUNT(DISTINCT SOURCE_ID) FROM #Table)
With the 6th insert commented out, should return A and B, with it included, should return A

Related

Recursive CTE to get a Category and all its ancestors [duplicate]

Given a child id, I need to return a query containing all parents of that child as well as their parents till I get to the root parent.
For example, given this data:
ID / Parent ID
1 / 0
2 / 1
3 / 2
4 / 0
5 / 3
So if I passed in ID 5 I would like to get a query with the results:
ID / Parent ID
1 / 0
2 / 1
3 / 2
This table does not work with a hierarchyid type so I suspect that this will need to be done with a CTE, but have no clue how. If it can be done in an SQL query / proc, any help would be appreciated.
Thanks
This is more or less what you want:
-- CTE to prepare hierarchical result set
;WITH #results AS
(
SELECT id,
parentid
FROM [table]
WHERE id = #childId
UNION ALL
SELECT t.id,
t.parentid
FROM [table] t
INNER JOIN #results r ON r.parentid = t.id
)
SELECT *
FROM #results;
Reference:
CTE: Common Table Expression
Working example:
-- create table with self lookup (parent id)
CREATE TABLE #tmp (id INT, parentid INT);
-- insert some test data
INSERT INTO #tmp (id, parentid)
SELECT 1,0 UNION ALL SELECT 2,1 UNION ALL SELECT 3,2
UNION ALL SELECT 4,0 UNION ALL SELECT 5,3;
-- prepare the child item to look up
DECLARE #childId INT;
SET #childId = 5;
-- build the CTE
WITH #results AS
(
SELECT id,
parentid
FROM #tmp
WHERE id = #childId
UNION ALL
SELECT t.id,
t.parentid
FROM #tmp t
INNER JOIN #results r ON r.parentid = t.id
)
-- output the results
SELECT *
FROM #results
WHERE id != #childId
ORDER BY id;
-- cleanup
DROP TABLE #tmp;
Output:
1 | 0
2 | 1
3 | 2

Remove Duplicates

I have a table like below:
SuppID AreaID SuppNo SupName SupPrice
------------------------------------------------
1 3 526 ANC 100
1 3 985 JTT 200
3 4 100 HIK 300
In the above table, for same SuppID(1) and same AreaID(3), different SuppNo are there (526 & 985) in two different rows.
In this scenario , I'd like to make those two rows into a single row with SuppNo field as blank.
Also my output result should display rows with all the columns.
Any Help?
This should get you started:
DECLARE #TABLE TABLE (SuppID INT, AreaID INT, SuppNo VARCHAR(5), SupName VARCHAR(5), SupPrice INT)
INSERT INTO #TABLE
SELECT 1,3,'526','ANC',100 UNION
SELECT 1,3,'985','JTT',200 UNION
SELECT 3,4,'100','HIK',300
-- select data before updates
SELECT * FROM #TABLE
-- add a row count by AreaID/SuppID
;WITH T1 AS
(
SELECT *
,SUM(1) OVER(PARTITION BY AREAID,SUPPID) AS ROWCNT
FROM #TABLE
)
-- set the SuppNo blank on rows that have more than 1 match
UPDATE T1 SET SuppNo='' WHERE ROWCNT>1
-- add a row # by AreaID/SuppID
;WITH T2 AS
(
SELECT *
,ROW_NUMBER() OVER(PARTITION BY AREAID,SUPPID ORDER BY AREAID,SUPPID) AS ROWID
FROM #TABLE
)
-- delete duplicate rows
DELETE
FROM T2
WHERE ROWID>1
-- select data after updates
SELECT * FROM #TABLE

Most effective way to get value if select count(*) = 1 with grouping

Lets say I have table with ID int, VALUE string:
ID | VALUE
1 abc
2 abc
3 def
4 abc
5 abc
6 abc
If I do select value, count(*) group by value I should get
VALUE | COUNT
abc 5
def 1
Now the tricky part, if there is count == 1 I need to get that ID from first table. Should I be using CTE? creating resultset where I will add ID string == null and run update b.ID = a.ID where count == 1 ?
Or is there another easier way?
EDIT:
I want to have result table like this:
ID VALUE count
null abc 5
3 def 1
If your ID values are unique, you can simply check to see if the max(id) = min(id). If so, then use either one, otherwise you can return null. Like this:
Select Case When Min(id) = Max(id) Then Min(id) Else Null End As Id,
Value, Count(*) As [Count]
From YourTable
Group By Value
Since you are already performing an aggregate, including the MIN and Max function is not likely to take any extra (noticeable) time. I encourage you to give this a try.
The way I would do it would indeed be a CTE:
using #group AS (SELECT value, Count(*) as count from MyTable GROUP BY value HAVING count = 1)
SELECT MyTable.ID, #group.value, #group.count from MyTable
JOIN #group ON #group.value = MyTable.value
When using group by, after the group by statement you can use a having clause.
So
SELECT [ID]
FROM table
GROUP BY [VALUE]
HAVING COUNT(*) = 1
Edit: with regards to your edited question: this uses some fun joins and unions
CREATE TABLE #table
(ID int IDENTITY,
VALUE varchar(3))
INSERT INTO #table (VALUE)
VALUES('abc'),('abc'),('def'),('abc'),('abc'),('abc')
SELECT * FROM (
SELECT Null as ID,VALUE, COUNT(*) as [Count]
FROM #table
GROUP BY VALUE
HAVING COUNT(*) > 1
UNION ALL
SELECT t.ID,t.VALUE,p.Count FROM
#table t
JOIN
(SELECT VALUE, COUNT(*) as [Count]
FROM #table
GROUP BY VALUE
HAVING COUNT(*) = 1) p
ON t.VALUE=p.VALUE
) a
DROP TABLE #table
maybe not the most efficient but something like this works:
SELECT MAX(Id) as ID,Value FROM Table WHERE COUNT(*) = 1 GROUP BY Value

one column split to more column sql server 2008?

Table name: Table1
id name
1 1-aaa-14 milan road
2 23-abcde-lsd road
3 2-mnbvcx-welcoome street
I want the result like this:
Id name name1 name2
1 1 aaa 14 milan road
2 23 abcde lsd road
3 2 mnbvcx welcoome street
This function ought to give you what you need.
--Drop Function Dbo.Part
Create Function Dbo.Part
(#Value Varchar(8000)
,#Part Int
,#Sep Char(1)='-'
)Returns Varchar(8000)
As Begin
Declare #Start Int
Declare #Finish Int
Set #Start=1
Set #Finish=CharIndex(#Sep,#Value,#Start)
While (#Part>1 And #Finish>0)Begin
Set #Start=#Finish+1
Set #Finish=CharIndex(#Sep,#Value,#Start)
Set #Part=#Part-1
End
If #Part>1 Set #Start=Len(#Value)+1 -- Not found
If #Finish=0 Set #Finish=Len(#Value)+1 -- Last token on line
Return SubString(#Value,#Start,#Finish-#Start)
End
Usage:
Select ID
,Dbo.Part(Name,1,Default)As Name
,Dbo.Part(Name,2,Default)As Name1
,Dbo.Part(Name,3,Default)As Name2
From Dbo.Table1
It's rather compute-intensive, so if Table1 is very long you ought to write the results to another table, which you could refresh from time to time (perhaps once a day, at night).
Better yet, you could create a trigger, which automatically updates Table2 whenever a change is made to Table1. Assuming that column ID is primary key:
Create Table Dbo.Table2(
ID Int Constraint PK_Table2 Primary Key,
Name Varchar(8000),
Name1 Varchar(8000),
Name2 Varchar(8000))
Create Trigger Trigger_Table1 on Dbo.Table1 After Insert,Update,Delete
As Begin
If (Select Count(*)From Deleted)>0
Delete From Dbo.Table2 Where ID=(Select ID From Deleted)
If (Select Count(*)From Inserted)>0
Insert Dbo.Table2(ID, Name, Name1, Name2)
Select ID
,Dbo.Part(Name,1,Default)
,Dbo.Part(Name,2,Default)
,Dbo.Part(Name,3,Default)
From Inserted
End
Now, do your data manipulation (Insert, Update, Delete) on Table1, but do your Select statements on Table2 instead.
The below solution uses a recursive CTE for splitting the strings, and PIVOT for displaying the parts in their own columns.
WITH Table1 (id, name) AS (
SELECT 1, '1-aaa-14 milan road' UNION ALL
SELECT 2, '23-abcde-lsd road' UNION ALL
SELECT 3, '2-mnbvcx-welcoome street'
),
cutpositions AS (
SELECT
id, name,
rownum = 1,
startpos = 1,
nextdash = CHARINDEX('-', name + '-')
FROM Table1
UNION ALL
SELECT
id, name,
rownum + 1,
nextdash + 1,
CHARINDEX('-', name + '-', nextdash + 1)
FROM cutpositions c
WHERE nextdash < LEN(name)
)
SELECT
id,
[1] AS name,
[2] AS name1,
[3] AS name2
/* add more columns here */
FROM (
SELECT
id, rownum,
part = SUBSTRING(name, startpos, nextdash - startpos)
FROM cutpositions
) s
PIVOT ( MAX(part) FOR rownum IN ([1], [2], [3] /* extend the list here */) ) x
Without additional modifications this query can split names consisting of up to 100 parts (that's the default maximum recursion depth, which can be changed), but can only display no more than 3 of them. You can easily extend it to however many parts you want it to display, just follow the instructions in the comments.
select T.id,
substring(T.Name, 1, D1.Pos-1) as Name,
substring(T.Name, D1.Pos+1, D2.Pos-D1.Pos-1) as Name1,
substring(T.Name, D2.Pos+1, len(T.name)) as Name2
from Table1 as T
cross apply (select charindex('-', T.Name, 1)) as D1(Pos)
cross apply (select charindex('-', T.Name, D1.Pos+1)) as D2(Pos)
Testing performance of suggested solutions
Setup:
create table Table1
(
id int identity primary key,
Name varchar(50)
)
go
insert into Table1
select '1-aaa-14 milan road' union all
select '23-abcde-lsd road' union all
select '2-mnbvcx-welcoome street'
go 10000
Result:
if you always will have 2 dashes, you can do the following by using PARSENAME
--testing table
CREATE TABLE #test(id INT, NAME VARCHAR(1000))
INSERT #test VALUES(1, '1-aaa-14 milan road')
INSERT #test VALUES(2, '23-abcde-lsd road')
INSERT #test VALUES(3, '2-mnbvcx-welcoome street')
SELECT id,PARSENAME(name,3) AS name,
PARSENAME(name,2) AS name1,
PARSENAME(name,1)AS name2
FROM (
SELECT id,REPLACE(NAME,'-','.') NAME
FROM #test)x
if you have dots in the name column you have to first replace them and then replace them back to dots in the end
example, by using a tilde to substitute the dot
INSERT #test VALUES(3, '5-mnbvcx-welcoome street.')
SELECT id,REPLACE(PARSENAME(name,3),'~','.') AS name,
REPLACE(PARSENAME(name,2),'~','.') AS name1,
REPLACE(PARSENAME(name,1),'~','.') AS name2
FROM (
SELECT id,REPLACE(REPLACE(NAME,'.','~'),'-','.') NAME
FROM #test)x

Finding duplicate rows but skip the last result?

I am trying to find duplicate rows in my DB, like this:
SELECT email, COUNT(emailid) AS NumOccurrences
FROM users
GROUP BY emailid HAVING ( COUNT(emailid) > 1 )
This returns the emailid and the number of matches found. Now what I want do is compare the ID column to another table I have and set a column there with the count.
The other table has a column named duplicates, which should contain the amount of duplicates from the select. So let's say we have 3 rows with the same emailid. The duplicates column has a "3" in all 3 rows. What I want is a "2" in the first 2 and nothing or 0 in the last of the 3 matching ID rows.
Is this possible?
Update:
I managed to have a temporary table now, which looks like this:
mailid | rowcount | AmountOfDups
643921 | 1 | 3
643921 | 2 | 3
643921 | 3 | 3
Now, how could I decide that only the first 2 should be updated (by mailid) in the other table? The other table has mailid as well.
SELECT ...
ROW_NUMBER() OVER (PARTITION BY email ORDER BY emailid DESC) AS RN
FROM ...
...is a great starting point for such a problem. Never underestimate the power of ROW_NUMBER()!
Using Sql Server 2005+ you could try something like (full example)
DECLARE #Table TABLE(
ID INT IDENTITY(1,1),
Email VARCHAR(20)
)
INSERT INTO #Table (Email) SELECT 'a'
INSERT INTO #Table (Email) SELECT 'b'
INSERT INTO #Table (Email) SELECT 'c'
INSERT INTO #Table (Email) SELECT 'a'
INSERT INTO #Table (Email) SELECT 'b'
INSERT INTO #Table (Email) SELECT 'a'
; WITH Duplicates AS (
SELECT Email,
COUNT(ID) TotalDuplicates
FROM #Table
GROUP BY Email
HAVING COUNT(ID) > 1
)
, Counts AS (
SELECT t.ID,
ROW_NUMBER() OVER(PARTITION BY t.Email ORDER BY t.ID) EmailID,
d.TotalDuplicates
FROM #Table t INNER JOIN
Duplicates d ON t.Email = d.Email
)
SELECT ID,
CASE
WHEN EmailID = TotalDuplicates
THEN 0
ELSE TotalDuplicates - 1
END Dups
FROM Counts