I trying to get the result table to contain rows where
Condition1: A.pc=B.pc AND
Condition2: (preferred (A.sub = B.Sub) or
else any one row that satisfy only Condition1)
I have tried the following inner join query and few other join and sub-query but can not figure out exact way to write a query with above strange condition.
SELECT * FROM tblA AS A INNER JOIN tblB AS B
ON A.sub=B.sub
WHERE A.pc=B.pc
tblA
-------------------
| id | pc | sub |
-------------------
| 0 | 5 | abc |
| 1 | 8 | def |
| 2 | 6 | ghi |
| 3 | 2 | jkl |
| 4 | 7 | mno |
| 5 | 19 | pqr |
-------------------
tblB
-------------------------
| pc | sub | uml | ull |
-------------------------
| 3 |arm | 1 | 1 |
| 3 |gtk | 1 | 2 |
| 3 |lmn | 1 | 3 |
| 3 |pop | 1 | 4 |
| 5 |abc | 1 | 5 |
| 5 |hlq | 1 | 6 |
| 5 |pon | 2 | 1 |
| 5 |qrt | 2 | 2 |
| 7 |alo | 2 | 3 |
| 7 |mno | 2 | 4 |
| 7 |ghm | 2 | 5 |
| 7 |stm | 2 | 6 |
| 9 |mck | 2 | 7 |
| 9 |plo | 3 | 1 |
| 9 |rtk | 3 | 2 |
| 9 |ert | 3 | 3 |
| 6 |gji | 3 | 4 |
| 6 |ghi | 3 | 5 |
| 6 |yux | 4 | 1 |
| 6 |del | 4 | 2 |
| 2 |jkl | 4 | 3 |
| 2 |jll | 5 | 4 |
| 2 |uin | 6 | 1 |
| 2 |tro | 6 | 2 |
| 19 |ppm | 6 | 3 |
| 19 |kde | 6 | 4 |
| 19 |grp | 6 | 5 |
| 19 |sho | 6 | 6 |
-------------------------
Expected Result Table:
-------------------------------
| id | pc | sub | uml | ull |
-------------------------------
| 0 | 5 |abc | 1 | 5 |
| 2 | 6 |ghi | 3 | 5 |
| 3 | 2 |jkl | 4 | 3 |
| 4 | 7 |mno | 2 | 4 |
| 5 | 19 |ppm | 6 | 3 | *
-------------------------------
* notice this is a arbitrary row as (A.sub=B.sub) not found
** notice there is no result for id=1 as pc=8 do not exist in tblB
Until someone comes up with a better answer, here is some code that does what you want.
Please, note it might not be a good solution in terms of performance (espcially as your tables grow).
SELECT *
FROM (
SELECT tblA.id, tblB.*
FROM tblA INNER JOIN tblB
ON tblA.pc = tblB.pc AND
tblA.id NOT IN (SELECT tblA.id
FROM tblA INNER JOIN tblB
ON tblA.sub = tblB.sub)
GROUP BY tblA.id
UNION
SELECT tblA.id, tblB.*
FROM tblA INNER JOIN tblB
ON tblA.sub = tblB.sub
GROUP BY tblA.id
) AS tu
ORDER BY id ASC;
See, also, this short demo.
One way of doing it I came up with is to repeat a join condition in where clause:
SELECT *
FROM tblA AS A
INNER JOIN tblB AS B
ON A.pc = B.pc
WHERE A.sub = B.sub
OR A.pc = B.pc
Related
I have the following problem:
Given the following table test in an oracle sql database:
+----+------+-------+------+
| id | name | start | stop |
+----+------+-------+------+
| 1 | A | 1 | 5 |
+----+------+-------+------+
| 2 | A | 2 | 6 |
+----+------+-------+------+
| 3 | A | 5 | 8 |
+----+------+-------+------+
| 4 | A | 9 | 10 |
+----+------+-------+------+
| 5 | B | 3 | 6 |
+----+------+-------+------+
| 6 | B | 4 | 8 |
+----+------+-------+------+
| 7 | B | 1 | 2 |
+----+------+-------+------+
I would like to find the number of overlapping intervals (endpoints included) [start, stop] n_overlap, for all id having the same name, i.e.:
+----+------+-------+------+-----------+
| id | name | start | stop | n_overlap |
+----+------+-------+------+-----------+
| 1 | A | 1 | 5 | 3 |
+----+------+-------+------+-----------+
| 2 | A | 2 | 6 | 3 |
+----+------+-------+------+-----------+
| 3 | A | 4 | 8 | 3 |
+----+------+-------+------+-----------+
| 4 | A | 9 | 10 | 1 |
+----+------+-------+------+-----------+
| 5 | B | 3 | 6 | 2 |
+----+------+-------+------+-----------+
| 6 | B | 4 | 8 | 2 |
+----+------+-------+------+-----------+
| 7 | B | 1 | 2 | 1 |
+----+------+-------+------+-----------+
One method uses a correlated subquery:
select t.*,
(select count(*)
from test t2
where t2.name = t.name and
t2.start < t.end and
t2.end > t.start
) as num_overlaps
from test t;
I have 3 tables:
+-----+---------+
|cl_id| name |
+-----+---------+
| 1 | adaf |
| 2 | rich | - clients
| 3 | call |
| 4 | alen |
| 5 | courney |
| 6 | warren |
+-----+---------+
+-----+---------+
|cl_id| data |
+-----+---------+
| 1 | 13 |
| 2 | 1000 | - table1
| 5 | 0 |
| 6 | 0 |
+-----+---------+
+-----+---------+
|cl_id| data |
+-----+---------+
| 2 | -355 | - table2
| 3 | 35 |
| 3 | 10 |
| 5 | 46 |
| 5 | 50 |
| 5 | 10 |
+-----+---------+
And I have to combine those three tables, so the result should be:
+-----+---------+--------+---------+
|cl_id| name |data_tb1|data_tb2 |
+-----+---------+--------+---------+
| 1 | adaf | 13 | 0 |
| 2 | rich | 1000 | -355 |
| 3 | call | 0 | 45 |
| 4 | alen | 0 | 0 |
| 5 | courney| 0 | 106 |
| 6 | warren | 0 | 0 |
+-----+---------+--------+---------+
It should output all clients and theirs SUM(data) from table1 and table2. clients goes one-to-more.
Thanks in advance
Simply using LEFT JOIN and GROUP BY
SELECT c.cl_id,
c.name,
COALESCE(SUM(t1.data), 0) AS data_tb1,
COALESCE(SUM(t2.data), 0) AS data_tb2
FROM clients c
LEFT JOIN table1 t1 ON c.cl_id = t1.cl_id
LEFT JOIN table2 t2 ON c.cl_id = t2.cl_id
GROUP BY c.cl_id,
c.name
ORDER BY c.cl_id;
If you are using SQL Server then simple Use Left Join as below :
SELECT C.cl_id,
C.name,
SUM(ISNULL(T.data, 0)) data_tb1,
SUM(ISNULL(T1.data, 0)) data_tb2
FROM
(
SELECT *
FROM clients
) C
LEFT JOIN table1 T ON T.cl_id = C.cl_id
LEFT JOIN table2 T1 ON T1.cl_id = C.cl_id
GROUP BY C.cl_id,
C.name
ORDER BY C.cl_id;
Desired Output :
+-----+---------+--------+---------+
|cl_id| name |data_tb1|data_tb2 |
+-----+---------+--------+---------+
| 1 | adaf | 13 | 0 |
| 2 | rich | 1000 | -355 |
| 3 | call | 0 | 45 |
| 4 | alen | 0 | 0 |
| 5 | courney| 0 | 106 |
| 6 | warren | 0 | 0 |
+-----+---------+--------+---------+
The scenario I am working on is as follows:-
A number of interviews are conducted a food poisoning case
A query called qryFoodInCase ( fldCaseID, fldFood) is generate consisting of all the food mentioned in all the interviews
An other query call qryFoodInInterview( fldCaseID, fldInterviewID, fldFood) consist of the food mentioned in each Interview
Now I am after the sql for a query that will return the food that was not consumed by an interviewee but consumed by one or more other interviewees.
The closest I've got is:
select Q1.fldCaseID,Q1.fldfood,Q2.fldInterviewID,fldGotSick
from qryFoodInCases as Q1
left join
(select * from qryFoodInInterview where qryFoodInInterview.fldInterviewID=1) as Q2
on Q1.fldFood=Q2.fldFood
where Q1.fldCaseID=1
The field Q2.fldInterviewID returns 1 for the food consumed and null for the food not consumed. However, I don't want to hard code the fldInterviewID in the sql. I would like a similar recordset returned for all the interviews in one query.
The SQL for qryFoodInCase and qryFoodInInterview are as follows:-
CREATE VIEW `qryFoodInCases`
AS
SELECT tblCases.fldCaseID
,fldfood
,count(tblFoodHistory.fldFoodID) AS fldFoodFrequency
FROM tblFood
INNER JOIN tblFoodHistory
ON tblFoodHistory.fldFoodID = tblFood.fldFoodID
INNER JOIN tblMealHistory
ON tblFoodHistory.fldMealID = tblMealHistory.fldMealHistoryID
INNER JOIN tblInterviews
ON tblInterviews.fldInterviewID = tblMealHistory.fldInterviewID
INNER JOIN tblCases
ON tblCases.fldCaseID = tblInterviews.fldCaseID
GROUP BY tblCases.fldCaseID, tblFood.fldFood
OUTPUT:
+-----------+------------+------------------+
| fldCaseID | fldFood | fldFoodFrequency |
+-----------+------------+------------------+
| 1 | Banana | 3 |
| 1 | Beans | 5 |
| 1 | Cabagge | 3 |
| 1 | Chicken | 1 |
| 1 | Pork | 5 |
| 1 | Potatoes | 1 |
| 1 | Rice | 1 |
| 1 | fried fish | 1 |
| 2 | Cabagge | 1 |
| 2 | Chicken | 2 |
| 2 | Potatoes | 1 |
| 2 | Rice | 1 |
| 2 | Salad | 1 |
+-----------+------------+------------------+
and
CREATE VIEW `qryFoodInInterview`
AS
SELECT tblInterviews.fldCaseID
,tblInterviews.fldInterviewID
,tblFood.fldFood
,tblInterviews.fldGotSick
FROM tblInterviews
INNER JOIN tblMealHistory
ON tblInterviews.fldInterviewID = tblMealHistory.fldInterviewID
INNER JOIN tblFoodHistory
ON tblFoodHistory.fldMealID = tblMealHistory.fldMealHistoryID
INNER JOIN tblFood
ON tblFood.fldFoodID = tblFoodHistory.fldFoodID
GROUP BY tblInterviews.fldInterviewID, tblFoodHistory.fldFoodID
OUTPUT
+-----------+----------------+------------+------------+
| fldCaseID | fldInterviewID | fldFood | fldGotSick |
+-----------+----------------+------------+------------+
| 1 | 1 | Pork | 0 |
| 1 | 1 | Banana | 0 |
| 1 | 1 | Rice | 0 |
| 1 | 1 | Potatoes | 0 |
| 1 | 2 | Chicken | 1 |
| 1 | 2 | Banana | 1 |
| 1 | 2 | Beans | 1 |
| 1 | 4 | Pork | 1 |
| 1 | 4 | fried fish | 1 |
| 1 | 4 | Beans | 1 |
| 2 | 6 | Salad | 0 |
| 2 | 6 | Chicken | 0 |
| 2 | 6 | Cabagge | 0 |
| 2 | 6 | Rice | 0 |
| 2 | 6 | Potatoes | 0 |
| 1 | 8 | Pork | 0 |
| 1 | 8 | Cabagge | 0 |
| 1 | 9 | Pork | 1 |
| 1 | 9 | Banana | 1 |
| 1 | 9 | Beans | 1 |
| 1 | 10 | Cabagge | 1 |
| 1 | 10 | Beans | 1 |
| 1 | 11 | Pork | 1 |
| 1 | 11 | Cabagge | 1 |
| 1 | 11 | Beans | 1 |
+-----------+----------------+------------+------------+
SQL Fiddle Demo
create a cross join off all food with all interviewrs
then a left join to see which one didnt had interview
the null mean interview didnt consume food
and exists mean someone different to interview consume food
.
SELECT F.fldFood, I.fldInterviewID, FI.fldInterviewID
FROM qryFoodInCases F, (SELECT DISTINCT fldInterviewID FROM qryFoodInInterview) I
LEFT JOIN qryFoodInInterview FI
ON F.fldFood = FI.fldFood
AND I.fldInterviewID = FI.fldInterviewID
WHERE FI.fldInterviewID IS NULL
AND EXISTS (SELECT 1
FROM qryFoodInInterview Q1
WHERE Q1.fldFood = F.fldFood
AND Q1.fldInterviewID <> I.fldInterviewID)
;
The trick was do the thing one step at the time. Once I realize I need all the combinations and create the cross join, the rest was easy.
select Q1.fldCaseID,Q1.fldfood,Q2.fldInterviewID,fldGotSick
from qryFoodInCases as Q1
left join
(select * from qryFoodInInterview QF1
JOIN qryFoodInInterview QF2
where QF1.fldInterviewID <> QF2.fldInterviewID
AND QF1.fldInterviewID > QF2.fldInterviewID
) as Q2
on Q1.fldFood=Q2.fldFood
where Q1.fldCaseID=1
As you mentioned you want to duplicate the items FoodInterviews, self JOIN will suffice your question. Please try this out , caution I didn't tested it.
Fixing Missed Self Join Condition to fix duplicates.
I already have the argument for the st column..
mysql> SELECT * FROM t ORDER BY LEFT(st,LOCATE(' ',st)), CAST(SUBSTRING(st,LOCATE(' ',st)+1) AS SIGNED);
+----+------+
| id | st |
+----+------+
| 1 | a 1 |
| 3 | a 6 |
| 4 | a 11 |
| 2 | a 11 |
| 5 | b 1 |
| 7 | b 6 |
| 8 | b 12 |
| 6 | b 12 |
+----+------+
this is what it should happen..
+----+------+
| id | st |
+----+------+
| 1 | a 1 |
| 3 | a 6 |
| 2 | a 11 |
| 4 | a 11 |
| 5 | b 1 |
| 7 | b 6 |
| 6 | b 12 |
| 8 | b 12 |
+----+------+
can I do this or not via SQL?
Well from what you show the only difference is that "ties" are broken by the ID column. If that's the case you can just use that as the third sort field:
SELECT *
FROM t
ORDER BY LEFT(st,LOCATE(' ',st)),
CAST(SUBSTRING(st,LOCATE(' ',st)+1) AS SIGNED),
ID
I am new to SQL, I need some help,
I want to find out confirmed (confirmed=1) reservation date with last name Steven
Can some one help for with the this SQL Query
table_member
+-----------+-----------+
|Member_id | last_name |
+-----------+-----------+
| 1 | David |
| 2 | owen |
| 3 | Tom |
| 4 | Steven |
| 5 | Steven |
| 6 | Steven |
| 7 | Steven |
| 8 | Steven |
| 9 | Steven |
+-----------+-----------+
table_resevation
+-----------+-----------+-----------+-----------+
|resvationID|Member_id | day | month |
+-----------+-----------+-----------+-----------+
| 1 | 1 | 1 | 6 |
| 2 | 1 | 2 | 6 |
| 3 | 2 | 1 | 6 |
| 4 | 4 | 1 | 6 |
| 5 | 4 | 2 | 6 |
| 6 | 5 | 1 | 6 |
| 7 | 5 | 2 | 6 |
| 8 | 7 | 3 | 6 |
+-----------+-----------+-----------+-----------+
Confirm
+-----------+-----------+-----------+
|confirmID |resvationID| confirmed |
+-----------+-----------+-----------+
| 1 | 1 | 1 |
| 2 | 2 | 1 |
| 3 | 3 | 0 |
| 4 | 4 | 0 |
| 5 | 5 | 1 |
| 6 | 6 | 1 |
| 7 | 7 | 0 |
| 8 | 8 | 1 |
+-----------+-----------+-----------+
Does something like this work for you?
SELECT
m.Member_ID, m.Last_Name,
r.Day, r.Month,
c.ConfirmID
FROM
table_member m
INNER JOIN
table_reservation r ON r.Member_ID = m.Member_ID
INNER JOIN
Confirm c ON c.resvationID = r.resvationID
WHERE
m.Last_Name = 'Steven' AND c.confirmed = 1
Select c.*, r.*, m.*
from confirm c
left join table_reservation r on c.reservationID = r.reservationID
left join table_member m on m.member_id = r.Member_id
Where c.confirmed = 1
and m.name = "Steven"
I did not test it, but that should point you to the way.
(Edit: second "where" was a typo)
Apply inner join on all table
select day,month from table_reservation inner join table_member on table_member.member_id=table_reservation.member_id
inner join confirm on confirm.reservationid=table_reservation.reservationid where confirm.confirmid=1
and table_member.last_name like 'Steven'