Reputation: 5398
I am new the SQL Server 2014. I have a table with records like the below.
Year | ItemName | price | effectivefromdate
===============================================
2018 | item27 | 1595.0000 | 2017-01-01
2018 | item27 | 1595.0000 | 2018-03-01
2018 | item29 | 1000.0000 | 2017-01-01
2018 | item29 | 1100.0000 | 2018-03-01
2018 | item30 | 1795.0000 | 2017-01-01
2018 | item30 | 1795.0000 | 2018-03-01
2018 | item30 | 1795.0000 | 2018-06-01
2018 | item32 | 1322.0000 | 2017-01-01
2018 | item32 | 1350.0000 | 2018-03-01
2018 | item32 | 1376.0000 | 2018-06-01
Here each item have one or more lines with same or different prices. I have to take the latest effective date for each item when the price has changed other wise If there is no price change with multiple effective dates then I have to return the item with minimum effective date.
For example, item27 have two entries but the price is not changed so I have to take price as 1595 and effective date as 2017-01-01 In case of item29, the price has changed here I have to take 1100 as price and effective date as 2018-03-01.
Expected Output
Year | ItemName | price | effectivefromdate
===============================================
2018 | item27 | 1595.0000 | 2017-01-01
2018 | item29 | 1100.0000 | 2018-03-01
2018 | item30 | 1795.0000 | 2017-01-01
2018 | item32 | 1376.0000 | 2018-06-01
I tried with Lag/Lead function but no luck. I am struggling for the past two days with this.
Please suggest me some solution to solve this.
Upvotes: 2
Views: 836
Reputation: 1855
By using Row_Number()
:
with cte as
(
Select Year, Itemname,price,effectivefromdate,
ROW_NUMBER() over (Partition by ItemName order by price desc, effectivefromdate asc) as ranking
from tbl
)
Select Year, Itemname,price,effectivefromdate from cte where ranking = 1
Note: This works only when price increases with time.
Upvotes: 3
Reputation: 17136
You can also using row_number and with group by like below. See a working demo here
; with cte as
(
select *, r= row_number() over( partition by ItemName order by effectivefromdate desc) from t
)
select
t1.Year,
t1.ItemName,
t1.Price,
effectivefromdate=min(t2.effectivefromdate) from cte t1 join
t t2 on r=1 and t1.Year=t2.Year
and t1.ItemName=t2.ItemName and t1.price=t2.price
group by
t1.Year,t1.ItemName,t1.Price
Upvotes: 2
Reputation: 1270463
You seem to want the effective date of the most recent price.
The idea is to get the set of rows that have the final price -- or said differently, that do not have a different price with a larger timestamp.
Then aggregate to get the earliest effective date:
select year, itemname, price, min(effectivefromdate)
from t
where not exists (select 1
from t t2
where t2.year = t.year and
t2.itemname = t.itemname and
t2.effectivefromdate > t.effectivefromdate and
t2.price <> t.price
)
group by year, itemname, price;
You can also approach this as a gaps-and-islands problem. However, this can be tricky -- particularly if prices can repeat with changes in-between.
Upvotes: 2