id city
1 London
2 Rome
3 London
4 Rome
Expected output like this:
London Rome
2 2
Using case expression...
How can I solve this query?
This is a standard aggregation + group by + having problem.
Sample data:
SQL> with test (id, city) as
2 (select 1, 'London' from dual union all
3 select 2, 'Rome' from dual union all
4 select 3, 'London' from dual union all
5 select 4, 'Rome' from dual union all
6 select 5, 'Zagreb' from dual
7 )
Query:
8 select city,
9 count(*) cnt
10 from test
11 group by city
12 having count(*) > 1;
CITY CNT
------ ----------
London 2
Rome 2
SQL>
What does case expression have to do with it?
You can get your output using conditional aggregation:
SELECT COUNT(CASE city WHEN 'London' THEN 1 END) AS London,
COUNT(CASE city WHEN 'Rome' THEN 1 END) AS Rome
FROM table_name
WHERE city IN ('London', 'Rome')
Which, for the sample data:
CREATE TABLE table_name (id, city) AS
SELECT 1, 'London' FROM DUAL UNION ALL
SELECT 2, 'Rome' FROM DUAL UNION ALL
SELECT 3, 'London' FROM DUAL UNION ALL
SELECT 4, 'Rome' FROM DUAL;
Outputs:
LONDON
ROME
2
2
However, in this case, it is unclear how to handle your requirement for repeated values if the values were not repeated.
fiddle
Related
I am trying to run this code in Oracle database but it's giving error :
ORA-00904: "vw_d"."cl_name": invalid identifier
What's wrong with the query:
SELECT *
FROM vw_doctrans vw_d
WHERE (SELECT COUNT(*)
FROM (SELECT *
FROM vw_doctrans vw
WHERE vw.cl_name = vw_d.cl_name
GROUP BY vw.country)) > 1
I tried this query in MySQL and works fine
To me, it looks like
Sample data:
SQL> with vw_doctrans (id, cl_name, country) as
2 (select 1, 'Bob', 'UK' from dual union all
3 select 2, 'Bob', 'USA' from dual union all
4 select 3, 'Bob', 'UAE' from dual union all
5 select 4, 'Bob', 'UAE' from dual union all
6 select 5, 'Bob', 'UAE' from dual union all
7 --
8 select 6, 'John', 'Canada' from dual union all
9 select 7, 'John', 'Canada' from dual union all
10 --
11 select 8, 'Caroline', 'India' from dual union all
12 select 9, 'Caroline', 'USA' from dual union all
13 select 10, 'Caroline', 'USA' from dual union all
14 select 11, 'Caroline', 'USA' from dual
15 ),
Query begins here:
16 temp as
17 (select v.*,
18 count(distinct country) over (partition by cl_name) cnt
19 from vw_doctrans v
20 )
21 select distinct cl_name, country
22 from temp
23 where cnt >= 2
24 order by cl_name, country;
CL_NAME COUNTR
-------- ------
Bob UAE
Bob UK
Bob USA
Caroline India
Caroline USA
SQL>
[EDIT: if the table already existed (i.e. without a CTE)]
Table contents:
SQL> select * from vw_doctrans;
ID CL_NAME COUNTR
---------- -------- ------
1 Bob UK
2 Bob USA
3 Bob UAE
4 Bob UAE
5 Bob UAE
6 John Canada
7 John Canada
8 Caroline India
9 Caroline USA
10 Caroline USA
11 Caroline USA
11 rows selected.
Here's how to use a CTE (that's line #16 in previous query):
SQL> with temp as
2 (select v.*,
3 count(distinct country) over (partition by cl_name) cnt
4 from vw_doctrans v
5 )
6 select distinct cl_name, country
7 from temp
8 where cnt >= 2
9 order by cl_name, country;
CL_NAME COUNTR
-------- ------
Bob UAE
Bob UK
Bob USA
Caroline India
Caroline USA
SQL>
You are way overthinking it. This is just:
SELECT vw_d.cl_name, vw_d.country
FROM vw_doctrans vw_d
WHERE EXISTS (
SELECT *
FROM vw_doctrans vw
WHERE vw.cl_name = vw_d.cl_name AND vw.country != vw_d.country
)
GROUP BY vw_d.cl_name, vw_d.country
But your "I need also other columns too: Bob with countries, date, message of text. Not only cl_name and country column" directly conflicts with your "if message came from same country 2-3times show only one time the country." You can't have both; if you want only one row for each name/country combination, you need to decide how to pick which values for the other columns you want.
I have 2 tables: SalesPeople and Customers that have snum and cnum as primary key respectively; both tables have city column as well.
Without using joins, we have to tell the names of customers and salespeople that belong to same city.
I have used nested queries to print the salespeople that belong to the city of customers, but cant figure out how to print customer names with this .
SELECT S.*
FROM SalesPeople S
WHERE City IN(
SELECT City
FROM Customers CX
);
How about this? (Disregard the fact that the WITH factoring clause doesn't exist in Oracle 9i (at least, I think so); you already have those tables).
Sample data:
SQL> with
2 salespeople (snum, city) as
3 (select 1, 'London' from dual union all
4 select 2, 'Paris' from dual union all
5 select 3, 'Rome' from dual
6 ),
7 customers (cnum, city) as
8 (select 100, 'Zagreb' from dual union all
9 select 101, 'Rome' from dual union all
10 select 102, 'Rome' from dual union all
11 select 103, 'Paris' from dual
12 )
Query:
13 select person_num
14 from (select snum as person_num, city from salespeople
15 union
16 select cnum, city from customers
17 )
18 where city = 'Rome';
PERSON_NUM
----------
3
101
102
SQL>
I have a sample table as below:
Target output:
I have tried with below script but only number of count produced as output:
select ID, count(COUNTRY) from test
group by COUNTRY
having count(COUNTRY)=3;
Could anyone here please help how can I get the targeted output? Thanks.
Here's one option - a subquery:
Sample data:
SQL> with test (id, name, country) as
2 (select 1, 'Mike' , 'Australia' from dual union all
3 select 2, 'Jason', 'Australia' from dual union all
4 select 3, 'Lee' , 'China' from dual union all
5 select 4, 'Simon', 'India' from dual union all
6 select 5, 'Alex' , 'Malaysia' from dual union all
7 select 6, 'John' , 'Australia' from dual
8 )
Query:
9 select id, country
10 from test
11 where country in (select country
12 from test
13 group by country
14 having count(*) = 3
15 )
16 order by id;
ID COUNTRY
---------- ---------
1 Australia
2 Australia
6 Australia
SQL>
You can use analytic functions and only scan the table once:
SELECT id, name, country
FROM (
SELECT id, name, country,
COUNT(*) OVER (PARTITION BY country) AS people_per_country
FROM table_name
)
WHERE people_per_country = 3;
Which, for the sample data:
CREATE TABLE table_name (id, name, country) AS
SELECT 1, 'Mike' , 'Australia' FROM DUAL UNION ALL
SELECT 2, 'Jason', 'Australia' FROM DUAL UNION ALL
SELECT 3, 'Lee' , 'China' FROM DUAL UNION ALL
SELECT 4, 'Simon', 'India' FROM DUAL UNION ALL
SELECT 5, 'Alex' , 'Malaysia' FROM DUAL UNION ALL
SELECT 6, 'John' , 'Australia' FROM DUAL;
Outputs:
ID
NAME
COUNTRY
1
Mike
Australia
2
Jason
Australia
6
John
Australia
db<>fiddle here
Sorry if I am not explaining this properly, I am relatively new to SQL.
In oracle have a table describing properties (city, property type, cost of rent per month, other information)
My question is: assuming 3 unique property types (hotel, house, empty lot), how can I show which cities do not have all 3 types of properties?
GROUP BY solution, make sure there are less than 3 different property types for each city returned:
select city
from tablename
where property_type in ('hotel', 'house', 'empty lot')
group by city
having count(distinct property_type) < 3
Your SQL query should be
SELECT City
FROM YourTable
WHERE hotel <> 'hotelname' and house <> 'housename' and emptylot <> 'name'
assuming
Hotel, House, Emptylot is column name in your database.
There are two ways,
GROUP BY
Analytic COUNT() OVER()
For example, Let's say I have sample data of 3 cities, where city 1 has all the property types satisfied, rest other cities are not having all the required property types.
Using GROUP BY
SQL> -- sample table data
SQL> WITH DATA AS(
2 SELECT 1 city, 'hotel' property FROM dual UNION ALL
3 SELECT 1 city, 'house' property FROM dual UNION ALL
4 SELECT 1 city, 'empty' property FROM dual UNION ALL
5 SELECT 2 city, 'hotel' property FROM dual UNION ALL
6 SELECT 2 city, 'house' property FROM dual UNION ALL
7 SELECT 2 city, 'scrap' property FROM dual UNION ALL
8 SELECT 3 city, 'empty' property FROM dual UNION ALL
9 select 3 city, 'house' property from dual
10 )
11 -- query
12 SELECT city
13 FROM data
14 WHERE property IN ('hotel', 'house', 'empty')
15 GROUP BY city
16 HAVING COUNT(property) < 3
17 /
CITY
----------
2
3
SQL>
Using Analytic COUNT() OVER()
SQL> -- sample table data
SQL> WITH DATA AS(
2 SELECT 1 city, 'hotel' property FROM dual UNION ALL
3 SELECT 1 city, 'house' property FROM dual UNION ALL
4 SELECT 1 city, 'empty' property FROM dual UNION ALL
5 SELECT 2 city, 'hotel' property FROM dual UNION ALL
6 SELECT 2 city, 'house' property FROM dual UNION ALL
7 SELECT 2 city, 'scrap' property FROM dual UNION ALL
8 SELECT 3 city, 'empty' property FROM dual UNION ALL
9 select 3 city, 'house' property from dual
10 )
11 -- query
12 SELECT DISTINCT city
13 FROM
14 (SELECT t.* ,
15 COUNT(property) OVER(PARTITION BY city ORDER BY city) rn
16 FROM DATA t
17 WHERE property IN ('hotel', 'house', 'empty')
18 )
19 WHERE rn < 3
20 /
CITY
----------
2
3
SQL>
Could be something like this:
SELECT City
FROM YourTable
WHERE [property type] != 'hotel' OR
[property type] != 'empty lot' OR
[property type] != 'house'
(Edited)Try this query :
select t.city from table_name t
where t.city NOT IN
(select city from table_name
where ( property_type ='hotel' or
property_type ='house' or
property_type ='Empty lot')
);
(Query returning cities where all three types of property are not present):
select t.city from table t inner join
(select city from table
where property_type not in ('House','Hotel','Empty lot')) x
on t.city=x.city
group by t.city
having count(*)<3 ;
I'm not 100% if tuples is the term for what I'm talking about but I'm looking at something like this:
Table grades
user grade
------------
Jim B
Bill C
Tim A
Jim B+
I know I can do:
SELECT COUNT(*)
FROM grades
WHERE (
(user = 'Jim' AND grade = 'B')
OR (user = 'Tim' AND grade = 'C')
);
But is there a way to do something more like this?
SELECT COUNT(*)
FROM grades
WHERE (user, grade) IN (('Jim','B'), ('Tim','C'));
EDIT: As a side note, I'd only tested with:
(user, grade) = ('Tim','C')
And that fails, so I assumed IN would fail as well, but I was wrong (thankfully!).
The query you posted should be valid syntax
SQL> ed
Wrote file afiedt.buf
1 with grades as (
2 select 'Jim' usr, 'B' grade from dual
3 union all
4 select 'Bill', 'C' from dual
5 union all
6 select 'Tim', 'A' from dual
7 union all
8 select 'Jim', 'B+' from dual
9 )
10 select *
11 from grades
12 where (usr,grade) in (('Jim','B'),
13 ('Tim','C'),
14* ('Tim','A'))
SQL> /
USR GR
---- --
Jim B
Tim A
You could use a subquery to treat a list of tuples like a table:
SELECT COUNT(*)
FROM grades
JOIN (
SELECT 'Jim' as user, 'B' as grade from dual
UNION ALL
SELECT 'Tim', 'C' from dual
UNION ALL
SELECT 'Pim', 'D' from dual
) as SearchTarget
ON SearchTarget.user = grades.user
and SearchTarget.grade = grades.grade