Hi apologies for formatting but im stumped and frustrated and i just need some help.
I've got two tables. I have made a good faith attempt to follow community standards but just in case it doesnt work, Table A has 3 columns 'ID', to identify a sales rep, 'Start' to indicate what company term they started, and 'Sales' to indicate their sales in that first term. Table B is just an expansion of Table A where it lists all terms (i marked it as quarters) a sales person was there and their sales.
Table A
+----+---------+-------+
| ID | Quarter | Sales |
+----+---------+-------+
| 1 | 141 | 30 |
| 2 | 151 | 50 |
| 3 | 151 | 80 |
+----+---------+-------+
Table B
+----+---------+-------+
| ID | Quarter | Sales |
+----+---------+-------+
| 1 | 141 | 30 |
| 1 | 142 | 25 |
| 1 | 143 | 45 |
| 2 | 151 | 50 |
| 2 | 152 | 60 |
| 2 | 153 | 75 |
| 3 | 151 | 80 |
| 3 | 152 | 50 |
| 3 | 153 | 70 |
+----+---------+-------+
My desired output is a table with ID, start term, sales from that term, second term, sales from that term, etc. for the first 6 terms an employee is there
my code is this
select a.id, start, a.sales,
case when a.start+1 = b.quarter then sales end as secondquartersales,
case when a.start+2 = b.quarter then sales end as thridquartersales,.....
from tablea a
left join tableb b
on a.id = b.id;
it gives nulls for all case when statements. please help
maybe try GROUP BY
create table a ( id number, strt number, sales number);
create table b (id number, quarter number , sales number);
insert into a values (1,141,30);
insert into a values (2,151,50);
insert into a values (3,151,80);
insert into b values ( 1,141,30);
insert into b values ( 1,142,25);
insert into b values ( 1,143,45);
insert into b values ( 2,151,50);
insert into b values ( 2,152,60);
insert into b values ( 2,153,75);
insert into b values ( 3,151,80);
insert into b values ( 3,152,50);
insert into b values ( 3,153,70);
select a.id, a.strt, a.sales,
max(case when a.strt+1 = b.quarter then b.sales end ) as secondquartersales,
max(case when a.strt+2 = b.quarter then b.sales end ) as thridquartersales
from a, b
where a.id = b.id
group by a.id, a.strt, a.sales;
OR PIVOT
select * from (
select a.id,
case when a.strt+1 = b.quarter then 'Q2'
when a.strt+2 = b.quarter then 'Q3'
when a.strt+3 = b.quarter then 'Q4'
when a.strt = b.quarter then 'Q1'end q,
b.sales sales
from a, b
where a.id = b.id)
pivot ( max(nvl(sales,0)) for Q in ('Q1', 'Q2', 'Q3', 'Q4'));
This is valid ANSI 92 SQL, as it is an inner join. The whole ANSI style version is just syntax candy.
Related
Below are two tables where there are parent bill number like 1, 4 and 8. These parents bill references to nothing/NULL values. They are referenced by one or more child bill number. For eg parent bill 1 is referenced by child bill 2, 3 and 6.
Table B also has the bill no column with prod code with actual service (ST values) and associated service values (SV). SV are the additional cost to ST.
Same ST may occur in multiple bill numbers. Here Bill number is only unique.
For eg, ST1 are in bill number 1 and 8. Also same SV may reference same or different ST.
SV1, SV2 and SV3 are referencing to ST1 corresponding to bill no. 1 and SV2 and SV4 are referencing to ST2 corresponding to bill no.2.
How can we get below expected output?
Table A:
| bill no | ref |
+----------------------------------------+
| 1 | |
| 2 | 1 |
| 3 | 1 |
| 4 | |
| 5 | 4 |
| 6 | 1 |
| 7 | 4 |
| 8 | |
| 9 | 8 |
Table B:
| bill no | Prod code | cost |
+-----------------------------------------------------+
| 1 | ST1 | 10
| 2 | SV1 | 20
| 3 | SV2 | 30
| 4 | ST2 | 10
| 5 | SV2 | 20
| 6 | SV3 | 30
| 7 | SV4 | 40
| 8 | ST1 | 50
| 9 | SV1 | 10
Expected output:
| bill no | Prod code | ST_cost | SV1 | SV2 | SV3 |
+---------------------------------------------------------------------------------------------+
| 1 | ST1 | 10 | 20 | 30 | 30 |
| 4 | ST2 | 10 | 20 | 40 | |
| 8 | ST1 | 50 | 10 | | |
Here's a script that should get you there:
USE tempdb;
GO
DROP TABLE IF EXISTS dbo.TableA;
CREATE TABLE dbo.TableA
(
BillNumber int NOT NULL PRIMARY KEY,
Reference int NULL
);
GO
INSERT dbo.TableA (BillNumber, Reference)
SELECT *
FROM (VALUES (1,NULL),
(2,1),
(3,1),
(4,NULL),
(5,4),
(6,1),
(7,4),
(8,NULL),
(9,8)) AS a(BillNumber, Reference);
GO
DROP TABLE IF EXISTS dbo.TableB;
CREATE TABLE dbo.TableB
(
BillNumber int NOT NULL PRIMARY KEY,
ProductCode varchar(10) NOT NULL,
Cost int NOT NULL
);
GO
INSERT dbo.TableB (BillNumber, ProductCode, Cost)
SELECT BillNumber, ProductCode, Cost
FROM (VALUES (1, 'ST1', 10),
(2, 'SV1', 20),
(3, 'SV2', 30),
(4, 'ST2', 10),
(5, 'SV2', 20),
(6, 'SV3', 30),
(7, 'SV4', 40),
(8, 'ST1', 50),
(9, 'SV1', 10)) AS b(BillNumber, ProductCode, Cost);
GO
WITH ParentBills
AS
(
SELECT b.BillNumber, b.ProductCode, b.Cost AS STCost
FROM dbo.TableB AS b
INNER JOIN dbo.TableA AS a
ON b.BillNumber = a.BillNumber
WHERE a.Reference IS NULL
),
SubBills
AS
(
SELECT pb.BillNumber, pb.ProductCode, pb.STCost,
b.ProductCode AS ChildProduct, b.Cost AS ChildCost
FROM ParentBills AS pb
INNER JOIN dbo.TableA AS a
ON a.Reference = pb.BillNumber
INNER JOIN dbo.TableB AS b
ON b.BillNumber = a.BillNumber
)
SELECT sb.BillNumber, sb.ProductCode, sb.STCost,
MAX(CASE WHEN sb.ChildProduct = 'SV1' THEN sb.ChildCost END) AS [SV1],
MAX(CASE WHEN sb.ChildProduct = 'SV2' THEN sb.ChildCost END) AS [SV2],
MAX(CASE WHEN sb.ChildProduct = 'SV3' THEN sb.ChildCost END) AS [SV3]
FROM SubBills AS sb
GROUP BY sb.BillNumber, sb.ProductCode, sb.STCost
ORDER BY sb.BillNumber;
You could write a function that creates you query based on your SV number.
And use "Execute Immediate" to execute the Query String and then "PIPE ROW" to generate the result.
Check This PIPE ROW EXAMPLE
I don't understand where the "SV1" value comes from on the second row.
But your problem is basically conditional aggregation:
with ab as (
select a.*, b.productcode, b.cost,
coalesce(a.reference, a.billnumber) as parent_billnumber
from a join
b
on b.billnumber = a.billnumber
)
select parent_billnumber,
max(case when reference is null then productcode end) as st,
sum(case when reference is null then cost end) as st_cost,
sum(case when productcode = 'SV1' then cost end) as sv1,
sum(case when productcode = 'SV2' then cost end) as sv2,
sum(case when productcode = 'SV3' then cost end) as sv3
from ab
group by parent_billnumber
order by parent_billnumber;
Here is a db<>fiddle.
Note this works because you have only one level of child relationships. If there are more, then recursive CTEs are needed. I would recommend that you ask a new question if this is possible.
The CTE doesn't actually add much to the query, so you can also write:
select coalesce(a.reference, a.billnumber) as parent_billnumber ,
max(case when a.reference is null then productcode end) as st,
sum(case when a.reference is null then b.cost end) as st_cost,
sum(case when b.productcode = 'SV1' then b.cost end) as sv1,
sum(case when b.productcode = 'SV2' then b.cost end) as sv2,
sum(case when b.productcode = 'SV3' then b.cost end) as sv3
from a join
b
on b.billnumber = a.billnumber
group by coalesce(a.reference, a.billnumber)
order by parent_billnumber;
I have the following tables
Table 1 : Product
id name
1 Bread
2 Bun
3 Cake
Table 2: Expense Items
product| quantity
1 | 100
2 | 150
3 | 180
1 | 25
2 | 30
Table 3: Income Items
product| quantity
1 | 100
2 | 150
3 | 180
1 | 25
2 | 30
Now I want the results like this
product | sum of quantity of expenseitem | sum of quantity of income item
1 | 125 | 125
2 | 180 | 180
3 | 180 | 180
What is the query to get this result ?
Thanks
You can try to use UNION ALL in a subquery with the condition in the aggregate function
Schema (PostgreSQL v9.6)
CREATE TABLE Product(
id int,
name varchar(50)
);
INSERT INTO Product VALUES (1,'Bread');
INSERT INTO Product VALUES (2,'Bun');
INSERT INTO Product VALUES (3,'Cake');
CREATE TABLE ExpenseItems(
product int,
quantity int
);
INSERT INTO ExpenseItems VALUES (1,100);
INSERT INTO ExpenseItems VALUES (2,150);
INSERT INTO ExpenseItems VALUES (3,180);
INSERT INTO ExpenseItems VALUES (1,25);
INSERT INTO ExpenseItems VALUES (2,30);
CREATE TABLE IncomeItems(
product int,
quantity int
);
INSERT INTO IncomeItems VALUES (1,100);
INSERT INTO IncomeItems VALUES (2,150);
INSERT INTO IncomeItems VALUES (3,180);
INSERT INTO IncomeItems VALUES (1,25);
INSERT INTO IncomeItems VALUES (2,30);
Query #1
SELECT p.id,
SUM(CASE WHEN grp = 1 THEN quantity END) SUMExpenseItems,
SUM(CASE WHEN grp = 2 THEN quantity END) SUMIncomeItems
FROM (
SELECT product, quantity,1 grp
FROM ExpenseItems
UNION ALL
SELECT product, quantity,2
FROM IncomeItems
) t1 JOIN Product p on p.id = t1.product
GROUP BY p.id;
| id | sumexpenseitems | sumincomeitems |
| --- | --------------- | -------------- |
| 1 | 125 | 125 |
| 2 | 180 | 180 |
| 3 | 180 | 180 |
View on DB Fiddle
Similar to #D-Shih's answer, PostgreSQL 9.4+ supports the FILTER() clause for conditional aggregation in place of CASE statements:
SELECT p.id,
SUM(quantity) FILTER (WHERE grp = 1) SUMExpenseItems,
SUM(quantity) FILTER (WHERE grp = 2) SUMIncomeItems
FROM
-- ...same union all query...
GROUP BY p.id
I have 2 tables
Table1
+---------+--------+-------+----+
| CALDATE | GROOMS | ROOMS | fn |
+---------+--------+-------+----+
| 1/5/18 | 15 | 17 | A12|
| 1/5/18 | 0 | 0 | A12|
| 1/6/18 | 0 | 0 | B34|
| 1/6/18 | 75 | 77 | B34|
| 1/7/18 | 123 | 125 | C56|
| 1/7/18 | 0 | 0 | C56|
+---------+--------+-------+----+
-
Table2
+----------+--------+----+
| ROOMDATE | pickup | FN |
+----------+--------+----+
| 1/5/18 | 0 | A12|
| 1/5/18 | 2 | A12|
| 1/5/18 | 1 | A12|
| 1/5/18 | 7 | A12|
| 1/6/18 | 2 | B34|
| 1/6/18 | 1 | B34|
| 1/6/18 | 13 | B34|
| 1/7/18 | 3 | C56|
| 1/7/18 | 0 | C56|
| 1/7/18 | 12 | C56|
+----------+--------+----+
Querying each I use
Select caldate as date, sum(grooms) as g, sum (rooms) as r
from Table1
and
Select roomdate as date, sum(pickup) as p
from Table2
These each give me the info I'm expecting, however when I try and join them things get wonky. I was hoping for something like
Select caldate as date,
sum(grooms) as g,
sum(rooms) as r,
sum(pickup) as p
from Table1
inner join table2 on table1.fn = table2.fn
But that returns way too high of each.
How do I join these queries so that I get my expected output of
+--------+-----+-----+----+----+
| Date | g | r | p | fn |
+--------+-----+-----+----+----+
| 1/5/18 | 15 | 17 | 10 | A12|
| 1/6/18 | 75 | 77 | 16 | B34|
| 1/7/18 | 123 | 125 | 15 | C56|
+--------+-----+-----+----+----+
Each row in your first table will match with each available row in the other table based on your join predicate. Take fn = A12 for example: since you have 2 rows in table1 and 4 rows in table2, you will end up with (4x2) 8 rows in your result set. That will cause your sums to be higher than they should be.
One way to fix this is to use derived tables to get your sums, then join them together:
SELECT t1.date, g, r, p, t1.fn
FROM (SELECT fn, caldate as date, sum(grooms) as g, sum (rooms) as r
FROM Table1
GROUP BY fn, caldate) t1
JOIN (SELECT fn, roomdate as date, sum(pickup) as p
FROM Table2
GROUP BY fn, roomdate) t2 on t1.fn = t2.fn
This makes sure there is one row returned from each table before the join.
You should be Grouping by 'caldate'. This way you will only be getting the sums per date.
Select caldate as date,
sum(grooms) as g,
sum(rooms) as r,
sum(pickup) as p
from Table1
inner join table2 on table1.fn = table2.fn
group by caldate
The reason that you get more rows than expected is because of join condition. if you want to find grooms and rooms for each fn on each day you have add date to join condition as well:
Select table1.caldate as date,
sum(grooms) as g,
sum(rooms) as r,
sum(pickup) as p
from table1
inner join table2 on table1.fn = table2.fn
and table1.caldate = table2.roomdate
Here's an answer that allows you to group by FN and date:
select format(caldate, 'M/d/yyyy') as date, sum(grooms) as g, sum(rooms) as r, t2.p, t1.fn
from Table1 t1
inner join (
select fn, roomdate, sum(pickup) as p from Table2 group by fn, roomdate
)t2 on t1.fn = t2.fn and t1.caldate = t2.roomdate
group by t1.caldate, t1.fn, t2.p
I created some sample data with DML statements so you can test grouping by different combinations of FN and date, and the output:
declare #t1 table (caldate datetime, grooms int, rooms int, fn varchar(3))
declare #t2 table (roomdate datetime, pickup int, fn varchar(3))
insert into #t1 select '1/5/18', 15, 17,'A12'
insert into #t1 select '1/5/18', 0, 0,'A12'
insert into #t1 select '1/6/18', 0, 0,'B34'
insert into #t1 select '1/6/18', 75, 77,'B34'
insert into #t1 select '1/7/18',123,125,'C56'
insert into #t1 select '1/8/18',100,200,'C56' -- changed to 1/8/18, changed vals
insert into #t2 select '1/5/18', 0 ,'A12'
insert into #t2 select '1/5/18', 2 ,'A12'
insert into #t2 select '1/5/18', 1 ,'A12'
insert into #t2 select '1/5/18', 7 ,'A12'
insert into #t2 select '1/6/18', 2 ,'B34'
insert into #t2 select '1/6/18', 1 ,'B34'
insert into #t2 select '1/6/18',13 ,'B34'
insert into #t2 select '1/7/18', 3 ,'C56'
insert into #t2 select '1/7/18', 0 ,'C56'
insert into #t2 select '1/8/18',12 ,'C56' -- changed to 1/8/18
select format(caldate, 'M/d/yyyy') as date, sum(grooms) as g, sum(rooms) as r, t2.p, t1.fn
from #t1 t1
inner join (
select fn, roomdate, sum(pickup) as p from #t2 group by fn, roomdate
)t2 on t1.fn = t2.fn and t1.caldate = t2.roomdate
group by t1.caldate, t1.fn, t2.p
Output:
date g r p fn
1/5/2018 15 17 10 A12
1/6/2018 75 77 16 B34
1/7/2018 123 125 3 C56
1/8/2018 100 200 12 C56
If you don't also join by date, then adding different combinations of caldate/fn give you duplicates. You must mean to join by the date as well, right?
I am trying to write something to automatically clean up some travel data. See these as flights:
FLIGHTS:
ID DocType Name Travel Date Fare Paid
1 INV Mrs G 13/03/2017 37.6
2 INV Mrs G 13/03/2017 200
3 INV Mr H 14/03/2017 60
4 INV Mr H 15/03/2017 126
5 CRN Mr H 15/03/2017 126
6 INV Mr H 20/03/2017 126
7 INV Mrs S 29/03/2017 110
8 INV Mr J 26/03/2017 54
9 INV Mr R 13/03/2017 200
10 INV Miss C 27/03/2017 78.98
Sometimes people buy a flight and then get a refund. This shows up as two identical entries in the data, except that the refund is DocType 'CRN'. I need to be able to pull both the booking and the refund line out of the dataset.
I can do this for the CRN tagged rows. But how can I pull out rows that are immediately above the CRN rows? The ID of the related INV row will always have an ID that is directly and sequentially lower than the CRN row.
I have managed
INSERT INTO TRAVEL.REFUNDS (ID, DocType, Name, [Travel Date], [Fare Paid])
SELECT ID, DocType, Name, [Travel Date], [Fare Paid]
FROM TRAVEL.FLIGHTS
WHERE [DocType] = 'CRN';
GO
Thank you in advance
using exists():
select *
from t
where DocType = 'CRN'
or exists (
select 1
from t i
where i.DocType='CRN'
and i.id-1 = t.id
)
or a left join
select t.*
from t
left join t i
on i.id-1 = t.id
where t.DocType = 'CRN'
or i.DocType = 'CRN'
rextester demo: rextester.com/MSGGX10058
returns:
+----+---------+--------+------------+----------+
| ID | DocType | Name | TravelDate | FarePaid |
+----+---------+--------+------------+----------+
| 4 | INV | Mr H | 15.03.2017 | 126.00 |
| 5 | CRN | Mr H | 15.03.2017 | 126.00 |
+----+---------+--------+------------+----------+
using not exists() for the opposite result set:
select *
from t
where DocType = 'INV'
and not exists (
select 1
from t i
where i.DocType='CRN'
and i.id-1 = t.id
)
returns:
+----+---------+--------+------------+----------+
| ID | DocType | Name | TravelDate | FarePaid |
+----+---------+--------+------------+----------+
| 1 | INV | Mrs G | 13.03.2017 | 37.60 |
| 2 | INV | Mrs G | 13.03.2017 | 200.00 |
| 3 | INV | Mr H | 14.03.2017 | 60.00 |
| 6 | INV | Mr H | 20.03.2017 | 126.00 |
| 7 | INV | Mrs S | 29.03.2017 | 110.00 |
| 8 | INV | Mr J | 26.03.2017 | 54.00 |
| 9 | INV | Mr R | 13.03.2017 | 200.00 |
| 10 | INV | Miss C | 27.03.2017 | 78.98 |
+----+---------+--------+------------+----------+
This is for SELECT purposes, not sure if you wanted that or INSERT or DELETE, but hopefully it's easily reworkable to those, plus it's good to check before modifying, right?
What I'm doing is, I'm using LAG/LEAD to add one new column, which is mostly a copy of some other row column, although shifted one row up or down. With that, you'll have each row containing everything needed to decide what to do with it, which will be done in the higher query that targets the lower query's results.
-- Making an MCVE, first time I know its name though.
DECLARE #Flights TABLE (ID int, DocType char(3))
INSERT INTO #Flights VALUES
( 1, 'INV')
, ( 2, 'INV')
, ( 3, 'INV') -- Should not show up.
, ( 4, 'CRN') -- Should not show up.
, ( 5, 'INV')
, ( 6, 'INV')
, ( 7, 'INV')
, ( 8, 'INV')
, ( 9, 'INV')
, (10, 'INV') -- Should not show up.
, (11, 'CRN') -- Should not show up.
-- Querying via LEAD(), with 1 level nesting (or subquerying, I dunno which is which).
SELECT *
FROM (
SELECT ID
, DocType AS DocTypeThis
, LEAD(DocType) OVER(ORDER BY ID ASC) AS DocTypeOther -- Seems like the choice of ASC/DESC reverses LAG/LEAD behaviours into each other, although not sure.
FROM #Flights
) AS T
WHERE (DocTypeOther IS NULL AND DocTypeThis = 'INV') -- Special treatment for last row (for other implementations, might be first row).
OR DocTypeThis = DocTypeOther -- This is the core of filtering, this fails only when the row is a 'CRN', or is superceded directly by a 'CRN'.
ORDER BY ID ASC
Table layout:
CREATE TABLE t_order (id INT, custId INT, order DATE)
I'm looking for a SQL command to select a maximum of one row per order (the customer who owns the order is identified by a field named custId).
I want to select ONE of the customer's orders (doesn't matter which one, say sorted by id) if there is no order date given for any of the rows.
I want to retrieve an empty Resultset for the customerId, if there is already a record with given order date.
Here is an example. Per customer there should be one order at most (one without a date given). Orders that have already a date value should not appear at all.
+---------------------------------------------------------+
|id | custId | date |
+---------------------------------------------------------+
| 1 10 NULL |
| 2 11 2008-11-11 |
| 3 12 2008-10-23 |
| 4 11 NULL |
| 5 13 NULL |
| 6 13 NULL |
+---------------------------------------------------------+
|
|
| Result
\ | /
\ /
+---------------------------------------------------------+
|id | custId | date |
+---------------------------------------------------------+
| 1 10 NULL |
| |
| |
| |
| 5 13 NULL |
| |
+---------------------------------------------------------+
powered be JavE
Edit:
I've choosen glavić's answer as the correct one, because it provides
the correct result with slightly modified data:
+---------------------------------------------------------+
|id | custId | date |
+---------------------------------------------------------+
| 1 10 NULL |
| 2 11 2008-11-11 |
| 3 12 2008-10-23 |
| 4 11 NULL |
| 5 13 NULL |
| 6 13 NULL |
| 7 11 NULL |
+---------------------------------------------------------+
Sfossen's answer will not work when customers appear more than twice because of its where clause constraint a.id != b.id.
Quassnoi's answer does not work for me, as I run server version 4.0.24 which yields the following error:
alt text http://img25.imageshack.us/img25/8186/picture1vyj.png
For a specific customer it's:
SELECT *
FROM t_order
WHERE date IS NULL AND custId=? LIMIT 1
For all customers its:
SELECT a.*
FROM t_order a
LEFT JOIN t_order b ON a.custId=b.custID and a.id != b.id
WHERE a.date IS NULL AND b.date IS NULL
GROUP BY custId;
Try this:
SELECT to1.*
FROM t_order AS to1
WHERE
to1.date IS NULL AND
to1.custId NOT IN (
SELECT to2.custId
FROM t_order AS to2
WHERE to2.date IS NOT NULL
GROUP BY to2.custId
)
GROUP BY to1.custId
For MySQL 4:
SELECT to1.*
FROM t_order AS to1
LEFT JOIN t_order AS to2 ON
to2.custId = to1.custId AND
to2.date IS NOT NULL
WHERE
to1.date IS NULL AND
to2.id IS NULL
GROUP BY to1.custId
This query will use one pass over index on custId.
For each distinct custId it will use one subquery over same index.
No GROUP BY, no TEMPORARY and no FILESORT — efficient, if your table is large.
SELECT VERSION()
--------
'4.1.22-standard'
CREATE INDEX ix_order_cust_id ON t_order(custId)
SELECT id, custId, order_date
FROM (
SELECT o.*,
CASE
WHEN custId <> #c THEN
(
SELECT 1
FROM t_order oi
WHERE oi.custId = o.custId
AND order_date IS NOT NULL
LIMIT 1
)
END AS n,
#c <> custId AS f,
#c := custId
FROM
(
SELECT #c := -1
) r,
t_order o
ORDER BY custId
) oo
WHERE n IS NULL AND f
---------
1, 10, ''
5, 13, ''
First filter out rows with dates, then filter out any row that has a similar row with a lower id. This should work because the matching record with the least id is unique if id is unique.
select * from t_order o1
where date is null
and not exists (select * from t_order o2
where o2.date is null
and o1.custId = o2.custId
and o1.id > o2.id)