Table with dates, table with week numbers, join together? - sql

Question

I have two tables. Table 1:
StuAp_Id StuAp_StaffID StuAp_Date StuAp_Attended
16 77000002659366 2011-09-07 Yes
17 77000002659366 2011-09-14 Yes
18 77000002659366 2011-09-14 Yes
19 77000002659366 2011-09-14 No
20 77000001171783 2011-09-19 Yes
Table 2:
Year Week Start
2011 1 2011-09-05 00:00:00.000
2011 2 2011-09-12 00:00:00.000
2011 3 2011-09-19 00:00:00.000
2011 4 2011-09-26 00:00:00.000
2011 5 2011-10-03 00:00:00.000
2011 6 2011-10-10 00:00:00.000
2011 7 2011-10-17 00:00:00.000
2011 8 2011-10-24 00:00:00.000
2011 9 2011-10-31 00:00:00.000
How would I join these two tables to make something like this:
StuAp_Id StuAp_StaffID StuAp_Date StuAp_Attended Week
16 77000002659366 2011-09-07 Yes 1
17 77000002659366 2011-09-14 Yes 2
18 77000002659366 2011-09-14 Yes 2
19 77000002659366 2011-09-14 No 2
20 77000001171783 2011-09-19 Yes 3
Thanks in advance

Answer 1

You can write simple INNER JOIN using GROUP BY clause.
SELECT Table1.*
,MAX(WEEK) AS WEEK
FROM Table1
INNER JOIN Table2 ON STUAP_DATE >= START
GROUP BY STUAP_ID,STUAP_STAFFID,STUAP_DATE,STUAP_ATTENDED

Answer 2

don't know about specifics on sql2k5 (don't have one around to test) but I would use a sub select eg.
select table_1.*,
[week] = (select isnull(max([week]), 0)
from table_2
where table_1.StuAp_Date >= table_2.start)
from table_1

Answer 3

CTEs to the rescue!
create table StuAp (
StuAp_Id int,
StuAp_StaffID bigint,
StuAp_Date datetime,
StuAp_Attended varchar(3)
)
create table Weeks (
Year int,
Week int,
Start datetime
)
insert into StuAp
values (16, 77000002659366, {d '2011-09-07'}, 'Yes'),
(17, 77000002659366, {d '2011-09-14'}, 'Yes'),
(18, 77000002659366, {d '2011-09-14'}, 'Yes'),
(19, 77000002659366, {d '2011-09-14'}, 'No'),
(20, 77000001171783, {d '2011-09-19'}, 'Yes')
insert into Weeks
values (2011, 1, {d '2011-09-05'}),
(2011, 2, {d '2011-09-12'}),
(2011, 3, {d '2011-09-19'}),
(2011, 4, {d '2011-09-26'}),
(2011, 5, {d '2011-10-03'}),
(2011, 6, {d '2011-10-10'}),
(2011, 7, {d '2011-10-17'}),
(2011, 8, {d '2011-10-24'}),
(2011, 9, {d '2011-10-31'})
;with OrderedWeeks as (
select ROW_NUMBER() OVER (ORDER BY year, week) as row, w.*
from Weeks w
), Ranges as (
select w1.*, w2.Start as Finish
from OrderedWeeks w1 inner join
OrderedWeeks w2 on w1.row = w2.row - 1
)
select s.StuAp_Id, s.StuAp_StaffID, s.StuAp_Date, s.StuAp_Attended, r.Week
from StuAp s inner join
Ranges r on s.StuAp_Date >= r.Start and s.StuAp_Date < r.Finish
This should scale quite well too.
Honestly though, if you find yourself doing queries like this often, you should really consider changing the stucture of your Weeks table to include a finish date. You could even make it an indexed view, or (assuming that the data changes rarely), you could keep your original table and use triggers or a SQL Agent job to keep a copy that contains Finish up to date.

Answer 4

SET ANSI_WARNINGS ON;
GO
DECLARE #Table1 TABLE
(
StuAp_Id INT PRIMARY KEY
,StuAp_StaffID NUMERIC(14,0) NOT NULL
,StuAp_Date DATETIME NOT NULL
,StuAp_Attended VARCHAR(3) NOT NULL
,StuAp_DateOnly AS DATEADD(DAY, DATEDIFF(DAY,0,StuAp_Date), 0) PERSISTED
);
INSERT #Table1
SELECT 16,77000002659366 ,'2011-09-07','Yes'
UNION ALL
SELECT 17,77000002659366 ,'2011-09-14','Yes'
UNION ALL
SELECT 18,77000002659366 ,'2011-09-14','Yes'
UNION ALL
SELECT 19,77000002659366 ,'2011-09-14','No'
UNION ALL
SELECT 20,77000001171783 ,'2011-09-19','Yes';
DECLARE #Table2 TABLE
(
Year INT NOT NULL
,Week INT NOT NULL
,Start DATETIME NOT NULL
,[End] AS DATEADD(DAY,6,Start) PERSISTED
,PRIMARY KEY(Year, Week)
,UNIQUE(Start)
);
INSERT #Table2
SELECT 2011,1 ,'2011-09-05 00:00:00.000'
UNION ALL
SELECT 2011,2 ,'2011-09-12 00:00:00.000'
UNION ALL
SELECT 2011,3 ,'2011-09-19 00:00:00.000'
UNION ALL
SELECT 2011,4 ,'2011-09-26 00:00:00.000'
UNION ALL
SELECT 2011,5 ,'2011-10-03 00:00:00.000'
UNION ALL
SELECT 2011,6 ,'2011-10-10 00:00:00.000'
UNION ALL
SELECT 2011,7 ,'2011-10-17 00:00:00.000'
UNION ALL
SELECT 2011,8 ,'2011-10-24 00:00:00.000'
UNION ALL
SELECT 2011,9 ,'2011-10-31 00:00:00.000';
--Solution 1 : if StuAp_Date has only date part
SELECT a.*, b.Week
FROM #Table1 a
INNER JOIN #Table2 b ON a.StuAp_Date BETWEEN b.Start AND b.[End]
--Solution 2 : if StuAp_Date has only date part
SELECT a.*, b.Week
FROM #Table1 a
INNER JOIN #Table2 b ON a.StuAp_Date BETWEEN b.Start AND DATEADD(DAY,6,b.Start)
--Solution 3 : if StuAp_Date has date & time
SELECT a.*, b.Week
FROM #Table1 a
INNER JOIN #Table2 b ON a.StuAp_DateOnly BETWEEN b.Start AND b.[End]
--Solution 4 : if StuAp_Date has date & time
SELECT a.*, b.Week
FROM #Table1 a
INNER JOIN #Table2 b ON DATEADD(DAY, DATEDIFF(DAY,0,a.StuAp_Date), 0) BETWEEN b.Start AND DATEADD(DAY,6,b.Start)