Assuming the following table:
ID Name Revision
--- ----- --------
1 blah 0
2 yada 1
3 blah 1
4 yada 0
5 blah 2
6 blah 3
How do I get the two records, one for "blah" and one for "yada" with highest revision number (3 for blah and 1 for yada)? Something like:
ID Name Revision
--- ----- --------
6 blah 3
2 yada 1
Also, once these records are retrieved, how do I get the rest, ordered by name and revision?
I am trying to create a master-detail view where master records are latest revisions and details include the previous revisions.
Basically, with the aggregate function MAX():
SELECT "Name", MAX("Revision") AS max_revison
FROM tbl
WHERE "Name" IN ('blah', 'yada');
GROUP BY "Name"
ORDER BY "Name"; -- ordering by revision would be pointless;
If you need more columns from the row, there are several ways. One would be to join the above subquery back to the base table:
SELECT t.*
FROM (
SELECT "Name", max("Revision") AS max_revison
FROM tbl
WHERE "Name" IN ('blah', 'yada');
GROUP BY "Name"
) AS sub
JOIN tbl AS t ON t."Revision" = sub.max_revison
AND t."Name" = sub."Name"
ORDER BY "Name";
Generally, this has the potential to yield more than one row per "Name" - if "Revision" is not unique (per "Name"). You would have to define how to pick one from a group of peers sharing the same maximum "Revision" - a tiebreaker.
Another way would be with NOT EXISTS, excluding rows that have greater peers, possibly faster:
SELECT t.*
FROM tbl AS t
WHERE "Name" IN ('blah', 'yada')
AND NOT EXISTS (
SELECT 1
FROM tbl AS t1
WHERE t1."Name" = t."Name"
AND t1."Revision" > t."Revision"
)
ORDER BY "Name";
Or you could use a CTE with an analytic function (window function):
WITH cte AS (
SELECT *, ROW_NUMBER() OVER(PARTITION BY "Name" ORDER BY "Revision" DESC) AS rn
FROM tbl
WHERE "Name" IN ('blah', 'yada')
)
SELECT *
FROM cte
WHERE rn = 1;
The last one is slightly different: one row per "Name" is guaranteed. If you don't use more ORDER BY items, an arbitrary row will be picked in case of a tie. If you want all peers use RANK() instead.
This approach will select the rows for each Name with the maximum revision number for that Name. The result will be the exact output you were looking for in your post.
SELECT *
FROM tbl a
WHERE a.revision = (select max(revision) from tbl where name = a.name)
ORDER BY a.name
In Oracle, you can use LAST function to simplify this.
select max(id) keep (dense_rank last order by revision),
name,
max(revision)
from table
group by name;
Demo
Related
I have the following data:
ID Date Num ClientID Dest
--------------------------------------------------------
123 04/29/2021 -2222 H1234 -1
123 04/29/2021 1 H1234 3
345 04/29/2021 -2222 H3456 -1
345 04/29/2021 1 H8888 .1
BTW: this does not include all the fields, just what I'm currently using for my query.
For every ID in the above table I'll always have 2 records. There are 2 scenarios that can take place:
As for ID = 123, the ClientID is the same
As for ID = 345, the ClientID is different
I'm trying to return the following data:
ID Date Num ClientID Dest
123 04/29/2021 1 H1234 3
The reason I'm returning only this row is because:
I only want 1 row per ID, where the ClientID is the same for both rows
Only need the record that does not have -2222, where the CLientID is the same for both rows
If the ClientID is different for the same ID (ex: 345), then completely skip these records.
Now the numbers for DEST can vary, so we can't always rely that one will be -1 and the other will be positive, however the NUM field will always have -2222 and 1 for the 2nd row (which is the row that I'd want to be returned)
I'm not sure how best to do this, I guess I thought about the alternative of just creating a CTE, and then counting the ClientID and if Count = 2 then select the data. The problem I find is with DEST field, I know that I can do Max(NUM) but since DEST field can very I wouldn't know how to select it.
Here is what I tried:
WITH Ranking AS (
SELECT Rank() OVER (PARTITION BY c.ID,c.date1,c.ClientID ORDER BY num
asc)x, c.*
FROM cte c
)
SELECT * FROM Ranking WHERE x = 2
I'm not if this is a good apprach, I guess it does the job but any thoughts?
One solution is to use the ever-useful window functions with row_number to select which pair to use and lead to check if both ClientIds are the same. This would work regardless if the specific values you have should change, plus is more performant than hitting the table twice:
select id, date, num, clientid, Dest from (
select *,
Row_Number() over(partition by id order by num) rn,
case when Lead(clientid) over(partition by id order by num)=clientid then 1 else 0 end same
from t
)t
where rn=1 and same=1
Just select all rows where Num isn't -2222 and the ID is in a subquery grouped by id and having only one distinct client id.
SELECT *
FROM tbl
WHERE ID IN (SELECT ID FROM tbl GROUP BY ID HAVING COUNT(DISTINCT ClientId) = 1)
AND Num != -2222
New to SQL here - I am trying to get 1 row from a table matching to a particular criteria
Typically this would look like
SELECT TOP 1 *
FROM myTable
WHERE id = 'abc'
The output may look like
value id
--------------
1 abc
The table has many entries for an 'id', and I am trying to get one entry per 'id'. Now I have list of 'id's. How would I execute something like
SELECT TOP 1 *
FROM myTable
FOR EACH id
WHERE id IN ('abc', 'edf', 'fgh')
Expecting result like
value id
--------------
1 abc
10 edf
12 fgh
I do not know if it is some sort union or concat operation, but would like to learn. I am working on Azure SQL Server
The table has many entries for an 'id', and I am trying to get one entry per 'id'. Now I have list of 'id's.
A typical method is row_number():
select t.*
from (select t.*,
row_number() over (partition by id order by id) as seqnum
from mytable t
) t
where seqnum = 1;
Note: you can filter on particular ids, if you want. It is unclear if that is really required for your question.
If you happen to be using SQL Server (as select top suggests), you can use the more concise, but somewhat less performant:
select top (1) with ties t.*
from mytable t
order by row_number() over (order by id order by (select null));
I want to find the max value in a column
ID CName Tot_Val PName
--------------------------------
1 1 100 P1
2 1 10 P2
3 2 50 P2
4 2 80 P1
Above is my table structure. I just want to find the max total value only from the table. In that four row ID 1 and 2 have same value in CName but total val and PName has different values. What I am expecting is have to find the max value in ID 1 and 2
Expected result:
ID CName Tot_Val PName
--------------------------------
1 1 100 P1
4 2 80 P1
I need result same as like mention above
select Max(Tot_Val), CName
from table1
where PName in ('P1', 'P2')
group by CName
This is query I have tried but my problem is that I am not able to bring PName in this table. If I add PName in the select list means it will showing the rows doubled e.g. Result is 100 rows but when I add PName in selected list and group by list it showing 600 rows. That is the problem.
Can someone please help me to resolve this.
One possible option is to use a subquery. Give each row a number within each CName group ordered by Tot_Val. Then select the rows with a row number equal to one.
select x.*
from ( select mt.ID,
mt.CName,
mt.Tot_Val,
mt.PName,
row_number() over(partition by mt.CName order by mt.Tot_Val desc) as No
from MyTable mt ) x
where x.No = 1;
An alternative would be to use a common table expression (CTE) instead of a subquery to isolate the first result set.
with x as
(
select mt.ID,
mt.CName,
mt.Tot_Val,
mt.PName,
row_number() over(partition by mt.CName order by mt.Tot_Val desc) as No
from MyTable mt
)
select x.*
from x
where x.No = 1;
See both solutions in action in this fiddle.
You can search top-n-per-group for this kind of a query.
There are two common ways to do it. The most efficient method depends on your indexes and data distribution and whether you already have another table with the list of all CName values.
Using ROW_NUMBER
WITH
CTE
AS
(
SELECT
ID, CName, Tot_Val, PName,
ROW_NUMBER() OVER (PARTITION BY CName ORDER BY Tot_Val DESC) AS rn
FROM table1
)
SELECT
ID, CName, Tot_Val, PName
FROM CTE
WHERE rn=1
;
Using CROSS APPLY
WITH
CTE
AS
(
SELECT CName
FROM table1
GROUP BY CName
)
SELECT
A.ID
,A.CName
,A.Tot_Val
,A.PName
FROM
CTE
CROSS APPLY
(
SELECT TOP(1)
table1.ID
,table1.CName
,table1.Tot_Val
,table1.PName
FROM table1
WHERE
table1.CName = CTE.CName
ORDER BY
table1.Tot_Val DESC
) AS A
;
See a very detailed answer on dba.se Retrieving n rows per group
, or here Get top 1 row of each group
.
CROSS APPLY might be as fast as a correlated subquery, but this often has very good performance (and better than ROW_NUMBER():
select t.*
from t
where t.tot_val = (select max(t2.tot_val)
from t t2
where t2.cname = t.cname
);
Note: The performance depends on having an index on (cname, tot_val).
I have the following temporary table
Aim is to flag the data with more than one records and put More than one records
In my example below, if Siren appears more than once, I would have
Siren ETS_RS Voie Ville nom_etp
348177155 POITOU-CHARENTES ENGRAIS P.C.E. (SNC) BOULEVARD WLADIMIR MORCH 17000 LA ROCHELLE More than one records
For records that are appearing once, I will have the single name of the company (here nom_etp)
Siren ETS_RS Voie Ville nom_etp
344843347 PRESTIGE AUTO ROCHELAIS (SAS) 4 RUE JEAN DEMEOCQ 17000 LA ROCHELLE NIGER
I tried a few things based on the idea that if I can have a count of more than one, I could flag them easily and use them with a CASE :
First: I tried to do a count
WITH cte_ssrep_moraux AS (...)
SELECT SIREN,ETS_RS,Voie,Ville
,Denomination AS nom_etp,COUNT(SIREN)
FROM cte_ssrep_moraux
GROUP BY ETS_RS,Voie,Ville,Denomination,SIREN
It hits a snitch as all counts were equal to one and I have the same dataset as in the picture...
Second:
WITH cte_ssrep_moraux AS (...)
SELECT ETS_RS,Voie,Ville
,Denomination AS nom_etp,SIREN,
RANK() OVER (PARTITION BY ETS_RS ORDER BY ETS_RS ASC) AS xx
FROM cte_ssrep_moraux
GROUP BY ETS_RS,Voie,Ville,Denomination,SIREN
It hits a snitch as all counts were equal to one and I have the same dataset as in the picture...
I'm bit confused on what I should do next. I have the feeling will be an easy one and I'll face palmed myself.
Many thanks for reading my question
If this is your criteria:
if Siren appears more than once,
Then the group by clause should only contain Siren:
SELECT SIREN, COUNT(*)
FROM cte_ssrep_moraux
GROUP BY SIREN
HAVING COUNT(*) > 1;
I'm not sure what you want to do after that, but this will return the SIREN values that appear more than once.
If there is more than one row and you change every nom_etp to 'more than one record', you end up with identical rows. That's why I prepared some tweaked query. See following (table simplified for clarity):
CREATE TABLE Duplicates
(
Id int,
Name varchar(20),
Item varchar(20)
)
INSERT Duplicates VALUES
(1,'Name1', 'Item1'),
(2,'Name2', 'Item2'),
(2,'Name2', 'Item3'),
(3,'Name3', 'Item4'),
(3,'Name3', 'Item5'),
(3,'Name3', 'Item6'),
(4,'Name4', 'Item7');
If you need just a query:
WITH Numbered AS
(
SELECT Id, Name, Item,
ROW_NUMBER() OVER (PARTITION BY Id ORDER BY Id) RowNum,
COUNT(*) OVER (PARTITION BY Id ORDER BY ID) TotalInGroup
FROM Duplicates
)
SELECT Id, Name,
CASE WHEN RowNum=1 AND TotalInGroup>1 THEN 'More records' ELSE Item END Item
FROM Numbered
If you need to normalize:
WITH Numbered AS
(
SELECT Id, Name, Item,
ROW_NUMBER() OVER (ORDER BY Id) Number,
ROW_NUMBER() OVER (PARTITION BY Id ORDER BY Id) RowNum,
COUNT(*) OVER (PARTITION BY Id ORDER BY ID) TotalInGroup
FROM Duplicates
)
MERGE Numbered AS tgt
USING Numbered AS src
ON src.Number=tgt.Number
WHEN MATCHED AND tgt.RowNum=1 AND tgt.TotalInGroup>1 THEN
UPDATE SET tgt.Item='More'
WHEN MATCHED AND tgt.RowNum>1 THEN
DELETE;
Table will look like below:
Id Name Item
-- ---- ----
1 Name1 Item1
2 Name2 More
3 Name3 More
4 Name4 Item7
If there are multiple rows with same id, first of them is updated with 'More' constant, all other in the group are deleted.
Use CTE for this purpose
;WITH CTE AS(
SELECT ETS_RS,Voie,Ville,Denomination AS nom_etp,SIREN,
ROW_NUMBER() OVER (PARTITION BY ETS_RS ORDER BY ETS_RS ASC) AS RN
FROM cte_ssrep_moraux
--GROUP BY ETS_RS,Voie,Ville,Denomination,SIREN
)
SELECT ETS_RS,
Voie,Ville,
CASE WHEN RN > 1 THEN 'More than one records'
ELSE nom_etp
END AS 'nom_etp',
SIREN
FROM CTE
;with cte
as
(
select siren,count(*) as cnt
from
yourtable
having count(*)>1
)
update t
set nom_etp='more than one records'
yourtable t where exists(Select 1 from cte c where c.sirenid=t.sirenid)
Since you still want all the records, including the unique.
Then you can use COUNT as a window function.
With a CASE to choose what to display as nom_etp.
select Siren, ETS_RS, Voie, Ville,
(case when count(*) over (partition by Siren) > 1 then 'More than one records' else nom_etp end) as nom_etp
from cte_ssrep_moraux;
Please find what I did
WITH cte_ssrep_moraux AS (
SELECT SIREN,ETS_RS,Voie,Ville
,Denomination AS nom_etp,ROW_NUMBER()
OVER (PARTITION BY ETS_RS ORDER BY ETS_RS ASC) AS Counting
FROM
(my_initial_cte) AS tb
)
SELECT Siren, ETS_RS, Voie, Ville,nom_etp
FROM cte_ssrep_moraux
WHERE counting = 1
AND Siren NOT IN (SELECT Siren FROM cte_ssrep_moraux WHERE counting > 1)
UNION ALL
SELECT DISTINCT Siren, ETS_RS, Voie, Ville,'More than one records'
FROM cte_ssrep_moraux
WHERE counting > 1
Explanation: After the initial CTE, I tried many of the solutions mentioned above especially using the CASE.
Issue with the CASE was that it would put something like that
Siren ETS_RS Voie Ville nom_etp
xxxx xyxy xyzet Bordeaux More than one records
xxxx xyxy xyzet Bordeaux More than one records
xxxx xyxy xyzet Bordeaux More than one records
xxxy zzzy ssare Paris Firm ABC
So instead of putting everything under a CASE, I said let's split that into 2 part :
First part would put everything with a counting equal to 1
Second part would put the rest with a counting that goes above 1 with a DISTINCT
Join the two results with an UNION ALL as the two sets have the same numbers of fetch rows
i have this query:
SELECT distinct
num as number,
name as name
from my_table_name
where number = '12345';
And this is the results:
number - name
1. 12345 - mike
2. 12345 - charlie
3. 12345 - jose
I need a new query when this happens (numbers duplicate or triplicate) show me only one of them. Example:
number - name
12345 - mike
I only need one of them; the position doesn't matter. If it find one, print this and close the procedure, function or cursor.
Distinct is going to return results that are distinct, relative to all of the data you are querying for. If you only want one of the results returned and you know that the result used is arbitrary, you can just add a filter based on the row number (how specifically this is done depends on what DBMS you are using.)
Oracle example:
select num as "number",
name as "name"
from my_table_name
where number = '12345'
and rownum = 1; -- just gets the first row.
SELECT * from
(SELECT rownum rnum,
num as number,
name as name
FROM my_table_name
WHERE number = '12345' )
WHERE rnum = 1
use ROW_NUMBER analytic function
SELECT *
FROM
(
select id, name, ROW_NUMBER() OVER ( partition by id order by name asc) as seq
from tableA
where number = '12345'
) T
where T.seq =1
If you don't care which one is being returned, why are you even asking for it to be returned?
However, to get a single line of results regardless of the number of matching rows, you should probably be using GROUP BY and a summary function:
Select
num as number,
max(name) as name --or min(), or any other summary function that works on this data type
from my_table_name
where num = '12345'
group by num