How to get rows from two tables on maximum value of particular field - sql

Question

I have two tables that has date_updated column.
TableA is like below
con_id date_updated type
--------------------------------------------
123 19/06/2018 2
123 15/06/2018 1
123 01/05/2018 3
101 06/04/2018 1
101 05/03/2018 2
And I have TableB that also has the same structure
con_id date_updated type
--------------------------------------------
123 15/05/2018 2
123 01/05/2018 1
101 07/06/2018 1
The resultant table should have the data with the recent date
con_id date_updated type
--------------------------------------------
123 19/06/2018 2
101 07/06/2018 1
Here the date_updated column is datetime datatype of sql server. I tried this by using group by and selecting the maximum date_updated. But i am not able to include column type in select statement. When i used type in group by ,the result is not correct as the type is also grouped. How can i query this. Please help

Answer 1

SELECT *
FROM
(SELECT *, ROW_NUMBER() OVER(Partition By con_id ORDER BY date_updated DESC) as seq
FROM
(SELECT * FROM TableA
UNION ALL
SELECT * FROM TableB) as tblMain) as tbl2
WHERE seq = 1

Answer 2

One method:
WITH A AS(
SELECT TOP 1 con_id,
date_updated,
type
FROM TableA
ORDER BY date_updated DESC),
B AS(
SELECT TOP 1 con_id,
date_updated,
type
FROM TableB
ORDER BY date_updated DESC),
U AS(
SELECT *
FROM A
UNION ALL
SELECT *
FROM B)
SELECT *
FROM U;
The 2 CTE's at the top get your most recent rows from the tables, and then the end statement unions them together.
For the benefit of the person who says this doesn't work:
USE Sandbox;
GO
CREATE TABLE tablea (con_id int, date_updated date, [type] tinyint);
CREATE TABLE tableb (con_id int, date_updated date, [type] tinyint);
GO
INSERT INTO tablea
VALUES
(123,'19/06/2018',2),
(123,'15/06/2018',1),
(123,'01/05/2018',3),
(101,'06/04/2018',1),
(101,'05/03/2018',2);
INSERT INTO tableb
VALUES
(123,'15/05/2018',2),
(123,'01/05/2018',1),
(101,'07/06/2018',1);
GO
WITH A AS(
SELECT TOP 1 con_id,
date_updated,
[type]
FROM TableA
ORDER BY date_updated DESC),
B AS(
SELECT TOP 1 con_id,
date_updated,
[type]
FROM TableB
ORDER BY date_updated DESC),
U AS(
SELECT *
FROM A
UNION ALL
SELECT *
FROM B)
SELECT *
FROM U;
GO
DROP TABLE tablea;
DROP TABLE tableb;
This returns the dataset:
con_id date_updated type
----------- ------------ ----
123 2018-06-19 2
101 2018-06-07 1
Which is identical to the OP's data:
con_id date_updated type
--------------------------------------------
123 19/06/2018 2
101 07/06/2018 1

Answer 3

Hope this helps:
WITH combined
AS(
select * FROM tableA
UNION
select * FROM tableB)
SELECT t1.con_id,
t1.date_updated,
t1.type
FROM (
SELECT con_id,
date_updated,
type,
row_number() OVER(partition BY con_id ORDER BY date_updated DESC) AS rownumber
FROM combined) t1
WHERE rownumber = 1;

Answer 4

Can be done using window functions:
declare #TableA table (con_id int, date_updated date, [type] int)
declare #TableB table (con_id int, date_updated date, [type] int)
insert into #TableA values
(123, '2018-06-19', 2)
, (123, '2018-06-15', 1)
, (123, '2018-05-01', 3)
, (101, '2018-04-06', 1)
, (101, '2018-03-05', 2)
insert into #TableB values
(123, '2018-05-15', 2)
, (123, '2018-05-01', 1)
, (101, '2018-06-07', 1)
select distinct con_id
, first_value(date_updated) over (partition by con_id order by con_id, date_updated desc) as con_id
, first_value([type]) over (partition by con_id order by con_id, date_updated desc) as [type]
from
(Select * from #TableA UNION Select * from #TableB) x