I have been trying all afternoon to try and achieve this with no success.
I have a db in with info on customers and the date that they purchase products from the store. It is grouped by a batch ID which I have converted into a date format.
So in my table I now have:
CustomerID|Date
1234 |2011-10-18
1234 |2011-10-22
1235 |2011-11-16
1235 |2011-11-17
What I want to achieve is to see the number of days between the most recent purchase and the last purchase and so on.
For example:
CustomerID|Date |Outcome
1234 |2011-10-18 |
1234 |2011-10-22 | 4
1235 |2011-11-16 |
1235 |2011-11-17 | 1
I have tried joining the table to itself but the problem I have is that I end up joining in the same format. I then tried with my join statement to return where it did <> match date.
Hope this makes sense, any help appreciated. I have searched all the relevant topics on here.
Will there be multiple groups of CustomerID? Or only and always grouped together?
DECLARE #myTable TABLE
(
CustomerID INT,
Date DATETIME
)
INSERT INTO #myTable
SELECT 1234, '2011-10-14' UNION ALL
SELECT 1234, '2011-10-18' UNION ALL
SELECT 1234, '2011-10-22' UNION ALL
SELECT 1234, '2011-10-26' UNION ALL
SELECT 1235, '2011-11-16' UNION ALL
SELECT 1235, '2011-11-17' UNION ALL
SELECT 1235, '2011-11-18' UNION ALL
SELECT 1235, '2011-11-19'
SELECT CustomerID,
MIN(date),
MAX(date),
DATEDIFF(day,MIN(date),MAX(date)) Outcome
FROM #myTable
GROUP BY CustomerID
SELECT a.CustomerID,
a.[Date],
ISNULL(DATEDIFF(DAY, b.[Date], a.[Date]),0) Outcome
FROM
(
SELECT ROW_NUMBER() OVER(PARTITION BY [CustomerID] ORDER BY date) Row,
CustomerID,
Date
FROM #myTable
) A
LEFT JOIN
(
SELECT ROW_NUMBER() OVER(PARTITION BY [CustomerID] ORDER BY date) Row,
CustomerID,
Date
FROM #myTable
) B ON a.CustomerID = b.CustomerID AND A.Row = B.Row + 1
Related
I am a SQL beginner. I have trouble on finding the answer of this question
For each customer_id who made an order on January 1, 2006, what was their historical (prior to January 1, 2006) first and last order dates?
I've tried to solve it using a subquery. But I don't know how to find the first and last order dates prior to Jan 1.
Columns of table A:
customer_id
order_id
order_date
revenue
product_id
Columns of table B:
product_id
category_id
SELECT customer_id, order_date FROM A
(
SELECT customer_id FROM A
WHERE order_date = ‘2006-01-01’
)
WHERE ...
There are two subqueries actually. First for "For each customer_id who made an order on January 1, 2006" and second for "their historical (prior to January 1, 2006) first and last order dates"
So, first:
select customer_id from A where order_date = '2006-01-01';
and second:
select customer_id, min(order_date) as first_date, max(order_date) as last_date
from A
where order_date < '2006-01-01' group by customer_id;
Finally you need to get only those customers from second subquery who exists in the first one:
select customer_id, min(order_date) as first_date, max(order_date) as last_date
from A as t1
where
order_date < '2006-01-01' and
customer_id in (
select customer_id from A where order_date = '2006-01-01')
group by customer_id;
or, could be more efficient:
select customer_id, min(order_date) as first_date, max(order_date) as last_date
from A as t1
where
order_date < '2006-01-01' and
exists (
select 1 from A as t2
where t1.customer_id = t2.customer_id and t2.order_date = '2006-01-01')
group by customer_id;
You can use conditionals in aggregate functions:
SELECT customer_id, MIN(order_date) AS first, MAX(order_date) AS last FROM A
WHERE customer_id IN (SELECT customer_id FROM A WHERE order_date = ‘2006-01-01’) AND order_date < '2006-01-01'
GROUP BY customer_id;
SELECT
CAST(c.DT AS DATE) AS 'Date'
, COUNT(p.PatternID) AS 'Count'
FROM CalendarMain c
LEFT OUTER JOIN Pattern p
ON c.DT = p.PatternDate
INNER JOIN Result r
ON p.PatternID = r.PatternID
INNER JOIN Detail d
ON p.PatternID = d.PatternID
WHERE r.Type = 7
AND d.Panel = 501
AND CAST(c.DT AS DATE)
BETWEEN '20190101' AND '20190201'
GROUP BY CAST(c.DT AS DATE)
ORDER BY CAST(c.DT AS DATE)
The query above isn't working for me. It still skips days where the COUNT is NULL for it's c.DT.
c.DT and p.PatternDate are both time DateTime, although c.DT can't be NULL. It is actually the PK for the table. It is populated as DateTimes for every single day from 2015 to 2049, so the records for those days exist.
Another weird thing I noticed is that nothing returns at all when I join C.DT = p.PatternDate without a CAST or CONVERT to a Date style. Not sure why when they are both DateTimes.
There are a few things to talk about here. At this stage it's not clear what you're actually trying to count. If it's the number of "patterns" per day for the month of Jan 2019, then:
Your BETWEEN will also count any activity occurring on 1 Feb,
It looks like one pattern could have multiple results, potentially causing a miscount
It looks like one pattern could have multiple details, potentially causing a miscount
If one pattern has say 3 eligible results, and also 4 details, you'll get the cross product of them. Your count will be 12.
I'm going to assume:
you only want the distinct number of patterns, regardless of the number of details and results.
You only want January's activity
--Set up some dummy data
DROP TABLE IF EXISTS #CalendarMain
SELECT cast('20190101' as datetime) as DT
INTO #CalendarMain
UNION ALL SELECT '20190102' as DT
UNION ALL SELECT '20190103' as DT
UNION ALL SELECT '20190104' as DT
UNION ALL SELECT '20190105' as DT --etc etc
;
DROP TABLE IF EXISTS #Pattern
SELECT cast('1'as int) as PatternID
,cast('20190101 13:00' as datetime) as PatternDate
INTO #Pattern
UNION ALL SELECT 2,'20190101 14:00'
UNION ALL SELECT 3,'20190101 15:00'
UNION ALL SELECT 4,'20190104 11:00'
UNION ALL SELECT 5,'20190104 14:00'
;
DROP TABLE IF EXISTS #Result
SELECT cast(100 as int) as ResultID
,cast(1 as int) as PatternID
,cast(7 as int) as [Type]
INTO #Result
UNION ALL SELECT 101,1,7
UNION ALL SELECT 102,1,8
UNION ALL SELECT 103,1,9
UNION ALL SELECT 104,2,8
UNION ALL SELECT 105,2,7
UNION ALL SELECT 106,3,7
UNION ALL SELECT 107,3,8
UNION ALL SELECT 108,4,7
UNION ALL SELECT 109,5,7
UNION ALL SELECT 110,5,8
;
DROP TABLE IF EXISTS #Detail
SELECT cast(201 as int) as DetailID
,cast(1 as int) as PatternID
,cast(501 as int) as Panel
INTO #Detail
UNION ALL SELECT 202,1,502
UNION ALL SELECT 203,1,503
UNION ALL SELECT 204,1,502
UNION ALL SELECT 205,1,502
UNION ALL SELECT 206,1,502
UNION ALL SELECT 207,2,502
UNION ALL SELECT 208,2,503
UNION ALL SELECT 209,2,502
UNION ALL SELECT 210,4,502
UNION ALL SELECT 211,4,501
;
-- create some variables
DECLARE #start_date as date = '20190101';
DECLARE #end_date as date = '20190201'; --I assume this is an exclusive end date
SELECT cal.DT
,isnull(patterns.[count],0) as [Count]
FROM #CalendarMain cal
LEFT JOIN ( SELECT cast(p.PatternDate as date) as PatternDate
,COUNT(DISTINCT p.PatternID) as [Count]
FROM #Pattern p
JOIN #Result r ON p.PatternID = r.PatternID
JOIN #Detail d ON p.PatternID = d.PatternID
WHERE r.[Type] = 7
and d.Panel = 501
GROUP BY cast(p.PatternDate as date)
) patterns ON cal.DT = patterns.patternDate
WHERE cal.DT >= #start_date
and cal.DT < #end_date --Your code would have included 1 Feb, which I assume was unintentional.
ORDER BY cal.DT
;
So I have a table as follows
product_id sender_id timestamp ...other columns...
1 2 1222
1 2 3423
1 2 1231
2 2 890
3 4 234
2 3 234234
I want to get rows where sender_id = 2, but I want to count and group by product_id and sort by timestamp descending. This means I need the following result
product_id sender_id timestamp count ...other columns...
1 2 3423 3
2 2 890 1
I tried the following query:
SELECT product_id, sender_id, timestamp, count(product_id), ...other columns...
FROM table
WHERE sender_id = 2
GROUP BY product_id
But I get the following error Error in query: ERROR: column "table.sender_id" must appear in the GROUP BY clause or be used in an aggregate function
Seems like I cannot SELECT columns that are not in the GROUP BY. Another method which I found online was to join
SELECT product_id, sender_id, timestamp, count, ...other columns...
FROM table
JOIN (
SELECT product_id, COUNT(product_id) AS count
FROM table
GROUP BY (product_id)
) table1 ON table.product_id = table1.product_id
WHERE sender_id = 2
GROUP BY product_id
But doing this simply lists all rows without grouping or counting. My guess is that the ON part simply extends table again.
Try grouping using product_id, sender_id
select product_id, sender_id, count(product_id), max(timestamp) maxtm
from t
where sender_id = 2
group by product_id, sender_id
order by maxtm desc
If you want other columns too:
select t.*, t1.product_count
from t
inner join (
select product_id, sender_id, count(product_id) product_count, max(timestamp) maxtm
from t
where sender_id = 2
group by product_id, sender_id
) t1
on t.product_id = t1.product_id and t.sender_id = t1.sender_id and t.timestamp = t1.maxtm
order by t1.maxtm desc
Just do a workout with your data:
CREATE TABLE products (product_id INTEGER,
sender_id INTEGER,
time_stamp INTEGER)
INSERT INTO products VALUES
(1,2,1222),
(1,2,3423),
(1,2,1231),
(2,2,890),
(3,4,234),
(2,3,234234)
SELECT product_id,sender_id,string_agg(time_stamp::text,','),count(product_id)
FROM products
WHERE sender_id=2
GROUP BY product_id,sender_id
Here you have distinct time_stamp ,so you need to apply some aggregate or just remove that column in select statement.
If you remove time_stamp in select statement then it would be very easy like below :
SELECT product_id,sender_id,count(product_id)
FROM products
WHERE sender_id=2
GROUP BY product_id,sender_id
I have following table
ID | Group | Type | Product
1 Dairy Milk Fresh Milk
2 Dairy Butter Butter Cream
3 Beverage Coke Coca cola
4 Beverage Diet Dew
5 Beverage Juice Fresh Juice
I need following output/query result:
ID | Group | Type | Product
1 Dairy
1 Milk Fresh Milk
2 Butter Butter Cream
2 Beverage
1 Coke Coca cola
2 Diet Dew
3 Juice Fresh Juice
For above sample a hard coded script can do the job but I look for a dynamic script for any number of groups. I do not have any idea how it can be done so, I do not have a sample query yet. I need ideas, examples that at least give me an idea. PIVOT looks a close option but does not looks to be fully fit for this case.
Here's a possible way. It basically unions the "Group-Headers" and the "Group-Items". The difficulty was to order them correctly.
WITH CTE AS
(
SELECT ID,[Group],Type,Product,
ROW_NUMBER() OVER (PARTITION BY [Group] Order By ID)AS RN
FROM Drink
)
SELECT ID,[Group],Type,Product
FROM(
SELECT RN AS ID,[Group],[Id]AS OriginalId,'' As Type,'' As Product, 0 AS RN, 'Group' As RowType
FROM CTE WHERE RN = 1
UNION ALL
SELECT RN AS ID,'' AS [Group],[Id]AS OriginalId,Type,Product, RN, 'Item' As RowType
FROM CTE
)X
ORDER BY OriginalId ASC
, CASE WHEN RowType='Group' THEN 0 ELSE 1 END ASC
, RN ASC
Here's a demo-fiddle: http://sqlfiddle.com/#!6/ed6ca/2/0
A slightly simplified approach:
With Groups As
(
Select Distinct Min(Id) As Id, [Group], '' As [Type], '' As Product
From dbo.Source
Group By [Group]
)
Select Coalesce(Cast(Z.Id As varchar(10)),'') As Id
, Coalesce(Z.[Group],'') As [Group]
, Z.[Type], Z.Product
From (
Select Id As Sort, Id, [Group], [Type], Product
From Groups
Union All
Select G.Id, Null, Null, S.[Type], S.Product
From dbo.Source As S
Join Groups As G
On G.[Group] = S.[Group]
) As Z
Order By Sort
It should be noted that the use of Coalesce is purely for aesthetic reasons. You could simply return null in these cases.
SQL Fiddle
And an approach with ROW_NUMBER:
IF OBJECT_ID('dbo.grouprows') IS NOT NULL DROP TABLE dbo.grouprows;
CREATE TABLE dbo.grouprows(
ID INT,
Grp NVARCHAR(MAX),
Type NVARCHAR(MAX),
Product NVARCHAR(MAX)
);
INSERT INTO dbo.grouprows VALUES
(1,'Dairy','Milk','Fresh Milk'),
(2,'Dairy','Butter','Butter Cream'),
(3,'Beverage','Coke','Coca cola'),
(4,'Beverage','Diet','Dew'),
(5,'Beverage','Juice','Fresh Juice');
SELECT
CASE WHEN gg = 0 THEN dr1 END GrpId,
CASE WHEN gg = 1 THEN rn1 END TypeId,
ISNULL(Grp,'')Grp,
CASE WHEN gg = 1 THEN Type ELSE '' END Type,
CASE WHEN gg = 1 THEN Product ELSE '' END Product
FROM(
SELECT *,
DENSE_RANK()OVER(ORDER BY Grp DESC) dr1
FROM(
SELECT *,
ROW_NUMBER()OVER(PARTITION BY Grp ORDER BY type,gg) rn1,
ROW_NUMBER()OVER(ORDER BY type,gg) rn0
FROM(
SELECT Grp,Type,Product, GROUPING(Grp) gg, GROUPING(type) tg FROM dbo.grouprows
GROUP BY Product, Type, Grp
WITH ROLLUP
)X1
WHERE tg = 0
)X2
WHERE gg=1 OR rn1 = 1
)X3
ORDER BY rn0
I have following data in the table as below and I am looking for a way to group the continuous time intervals for each id to return:
CREATE TABLE DUMMY
(
ID VARCHAR2(10 BYTE),
TIME_STAMP VARCHAR2(8 BYTE),
NAME VARCHAR2(255 BYTE)
);
SELECT ID, min(TIME_STAMP) "startDate", max(TIME_STAMP) "endDate", NAME
GROUP BY ID , NAME
something like
100 20011128 20011203 David
100 20011204 20011207 Unknown
100 20011208 20011215 David
100 20011216 20011220 Sara
and so on ...
ps. I have a sample script, but i don't know how to attach my file.
Hi every one here is more input:
There is only one record with time_stamp for a specific ID.
Users can be different, for example for day 1 David, day 2 unknown, day 3 David and so on.
So there is one row for every day of year for each ID but with different users.
Now, i want to see the break point, differences base on time_stamp intervals from day one
until last day for a specific ID in day order from begin day until last day.
Query Result should be :
ID NAME MIN_DATE MAX_DATE
100 David 20011128 20050407
100 Sara 20050408 20050417
100 David 20050418 20080416
100 Unknown 20080417 20080507
100 David 20080508 20080508
100 Unknown 20080509 20080607
100 David 20080608 20080608
100 Unknown 20080609 20080921
100 David 20080922 20080922
100 Unknown 20080923 20081231
100 David 20090101 20090405
thanks
Hi again, many thanks to everyone, i have solved the problem, here is the solution:
select id, min(time_stamp), max(time_stamp), name
from ( select id, time_stamp, name,
max(rn) over (order by time_stamp) grp
from ( select id, time_stamp, name,
case
when lag(name) over (order by time_stamp) <> name or
row_number() over (order by time_stamp) = 1
then row_number() over (order by time_stamp)
end rn
from dummy
)
)
group by id, grp, name
order by 1
Select
ID,
Name,
min(time_stamp) min_date,
max(time_stamp) max_date
from
Dummy
group by
Id,
Name
That should work.
IF you want the date range for each Id, but all the names you can do:
Select
d.Id,
d.Name,
dr.min_date,
dr.max_date
from
Dummy d
JOIN
(Select
Id,
min(time_stamp) min_date,
max(time_stamp) max_date
from
Dummy
group by
Id
) dr
on ( dr.Id = d.Id)