Select top three values in each group - postgresql

following is my sample table and rows
create table com (company text,val int);
insert into com values ('com1',1),('com1',2),('com1',3),('com1',4),('com1',5);
insert into com values ('com2',11),('com2',22),('com2',33),('com2',44),('com2',55);
insert into com values ('com3',111),('com3',222),('com3',333),('com3',444),('com3',555);
I want to get the top 3 value of each company, expected output is :
company val
---------------
com1 5
com1 4
com1 3
com2 55
com2 44
com2 33
com3 555
com3 444
com3 333

Try This:
SELECT company, val FROM
(
SELECT *, ROW_NUMBER() OVER (PARTITION BY
company order by val DESC) AS Row_ID FROM com
) AS A
WHERE Row_ID < 4 ORDER BY company
--Quick Demo Here...

Since v9.3 you can do a lateral join
select distinct com_outer.company, com_top.val from com com_outer
join lateral (
select * from com com_inner
where com_inner.company = com_outer.company
order by com_inner.val desc
limit 3
) com_top on true
order by com_outer.company;
It might be faster but, of course, you should test performance specifically on your data and use case.

You can try arrays, which are available since Postgres v9.0.
WITH com_ordered AS (SELECT * FROM com ORDER BY company,val DESC)
SELECT company,unnest((array_agg(val))[0:3])
FROM com_ordered GROUP BY company;

Related

PostgreSQL select different columns based on condition [duplicate]

I have two queries :
Queries Simplified excluding Joins
Query 1 : select ProductName,NumberofProducts (in inventory) from Table1.....;
Query 2 : select ProductName, NumberofProductssold from Table2......;
I would like to know how I can get an output as :
ProductName NumberofProducts(in inventory) ProductName NumberofProductsSold
The relationships used for getting the outputs for each query are different.
I need the output this way for my SSRS report .
(I tried the union statement but it doesnt work for the output I want to see. )
Here is an example that does a union between two completely unrelated tables: the Student and the Products table. It generates an output that is 4 columns:
select
FirstName as Column1,
LastName as Column2,
email as Column3,
null as Column4
from
Student
union
select
ProductName as Column1,
QuantityPerUnit as Column2,
null as Column3,
UnitsInStock as Column4
from
Products
Obviously you'll tweak this for your own environment...
I think you are after something like this; (Using row_number() with CTE and performing a FULL OUTER JOIN )
Fiddle example
;with t1 as (
select col1,col2, row_number() over (order by col1) rn
from table1
),
t2 as (
select col3,col4, row_number() over (order by col3) rn
from table2
)
select col1,col2,col3,col4
from t1 full outer join t2 on t1.rn = t2.rn
Tables and data :
create table table1 (col1 int, col2 int)
create table table2 (col3 int, col4 int)
insert into table1 values
(1,2),(3,4)
insert into table2 values
(10,11),(30,40),(50,60)
Results :
| COL1 | COL2 | COL3 | COL4 |
---------------------------------
| 1 | 2 | 10 | 11 |
| 3 | 4 | 30 | 40 |
| (null) | (null) | 50 | 60 |
How about,
select
col1,
col2,
null col3,
null col4
from Table1
union all
select
null col1,
null col2,
col4 col3,
col5 col4
from Table2;
The problem is that unless your tables are related you can't determine how to join them, so you'd have to arbitrarily join them, resulting in a cartesian product:
select Table1.col1, Table1.col2, Table2.col3, Table2.col4
from Table1
cross join Table2
If you had, for example, the following data:
col1 col2
a 1
b 2
col3 col4
y 98
z 99
You would end up with the following:
col1 col2 col3 col4
a 1 y 98
a 1 z 99
b 2 y 98
b 2 z 99
Is this what you're looking for? If not, and you have some means of relating the tables, then you'd need to include that in joining the two tables together, e.g.:
select Table1.col1, Table1.col2, Table2.col3, Table2.col4
from Table1
inner join Table2
on Table1.JoiningField = Table2.JoiningField
That would pull things together for you into however the data is related, giving you your result.
If you mean that both ProductName fields are to have the same value, then:
SELECT a.ProductName,a.NumberofProducts,b.ProductName,b.NumberofProductsSold FROM Table1 a, Table2 b WHERE a.ProductName=b.ProductName;
Or, if you want the ProductName column to be displayed only once,
SELECT a.ProductName,a.NumberofProducts,b.NumberofProductsSold FROM Table1 a, Table2 b WHERE a.ProductName=b.ProductName;
Otherwise,if any row of Table1 can be associated with any row from Table2 (even though I really wonder why anyone'd want to do that), you could give this a look.
Old question, but where others use JOIN to combine unrelated queries to rows in one table, this is my solution to combine unrelated queries to one row, e.g:
select
(select count(*) c from v$session where program = 'w3wp.exe') w3wp,
(select count(*) c from v$session) total,
sysdate
from dual;
which gives the following one-row output:
W3WP TOTAL SYSDATE
----- ----- -------------------
14 290 2020/02/18 10:45:07
(which tells me that our web server currently uses 14 Oracle sessions out of the total of 290 sessions; I log this output without headers in an sqlplus script that runs every so many minutes)
Load each query into a datatable:
http://www.dotnetcurry.com/ShowArticle.aspx?ID=143
load both datatables into the dataset:
http://msdn.microsoft.com/en-us/library/aeskbwf7%28v=vs.80%29.aspx
This is what you can do. Assuming that your ProductName column have common values.
SELECT
Table1.ProductName,
Table1.NumberofProducts,
Table2.ProductName,
Table2.NumberofProductssold
FROM Table1
INNER JOIN Table2
ON Table1.ProductName= Table2.ProductName
Try this:
SELECT ProductName,NumberofProducts ,NumberofProductssold
FROM table1
JOIN table2
ON table1.ProductName = table2.ProductName
Try this:
GET THE RECORD FOR CURRENT_MONTH, LAST_MONTH AND ALL_TIME AND MERGE THEM INTO SINGLE ARRAY
$analyticsData = $this->user->getMemberInfoCurrentMonth($userId);
$analyticsData1 = $this->user->getMemberInfoLastMonth($userId);
$analyticsData2 = $this->user->getMemberInfAllTime($userId);
foreach ($analyticsData2 as $arr) {
foreach ($analyticsData1 as $arr1) {
if ($arr->fullname == $arr1->fullname) {
$arr->last_send_count = $arr1->last_send_count;
break;
}else{
$arr->last_send_count = 0;
}
}
foreach ($analyticsData as $arr2) {
if ($arr->fullname == $arr2->fullname) {
$arr->current_send_count = $arr2->current_send_count;
break;
}else{
$arr->current_send_count = 0;
}
}
}
echo "<pre>";
print_r($analyticsData2);die;

find set of values that have max occurrence in column in postgresql

Print the name(s) and sid(s) of the student(s) enrolled in the most classes
Enroll
sid class number
1 23
2 54
1 54
3 43
1 43
2 43
student
sid sname
1 sagar
2 kiran
3 ravi
4 vishal
output
sid sname
1 sagar
Group enrollments by students, order by count and use limit 1:
select s.id, s.name
from student s
join enroll e on e.sid = s.id
group by s.id, s.name
order by count(*) desc
limit 1
Note how you don't need the select count(*) - you may simply refer to it.
I think this will help you
SELECT <column_name> FROM <table_name> WHERE <column_name>=
(SELECT <column_name>
FROM (SELECT <column_name>, count(*) as cnt FROM <table_name> GROUP BY <column_name>) AS foo
WHERE foo.cnt=(SELECT MAX(c) FROM (SELECT <column_name>,count(*) AS c FROM <column_name> GROUP BY <column_name>) AS bar)) limit 1

How to rank in postgres query

I'm trying to rank a subset of data within a table but I think I am doing something wrong. I cannot find much information about the rank() feature for postgres, maybe I'm looking in the wrong place. Either way:
I'd like to know the rank of an id that falls within a cluster of a table based on a date. My query is as follows:
select cluster_id,feed_id,pub_date,rank
from (select feed_id,pub_date,cluster_id,rank()
over (order by pub_date asc) from url_info)
as bar where cluster_id = 9876 and feed_id = 1234;
I'm modeling this after the following stackoverflow post: postgres rank
The reason I think I am doing something wrong is that there are only 39 rows in url_info that are in cluster_id 9876 and this query ran for 10 minutes and never came back. (actually re-ran it for quite a while and it returned no results, yet there is a row in cluster 9876 for id 1234) I'm expecting this will tell me something like "id 1234 was 5th for the criteria given). It will return a relative rank according to my query constraints, correct?
This is postgres 8.4 btw.
By placing the rank() function in the subselect and not specifying a PARTITION BY in the over clause or any predicate in that subselect, your query is asking to produce a rank over the entire url_info table ordered by pub_date. This is likely why it ran so long as to rank over all of url_info, Pg must sort the entire table by pub_date, which will take a while if the table is very large.
It appears you want to generate a rank for just the set of records selected by the where clause, in which case, all you need do is eliminate the subselect and the rank function is implicitly over the set of records matching that predicate.
select
cluster_id
,feed_id
,pub_date
,rank() over (order by pub_date asc) as rank
from url_info
where cluster_id = 9876 and feed_id = 1234;
If what you really wanted was the rank within the cluster, regardless of the feed_id, you can rank in a subselect which filters to that cluster:
select ranked.*
from (
select
cluster_id
,feed_id
,pub_date
,rank() over (order by pub_date asc) as rank
from url_info
where cluster_id = 9876
) as ranked
where feed_id = 1234;
Sharing another example of DENSE_RANK() of PostgreSQL.
Find top 3 students sample query.
Reference taken from this blog:
Create a table with sample data:
CREATE TABLE tbl_Students
(
StudID INT
,StudName CHARACTER VARYING
,TotalMark INT
);
INSERT INTO tbl_Students
VALUES
(1,'Anvesh',88),(2,'Neevan',78)
,(3,'Roy',90),(4,'Mahi',88)
,(5,'Maria',81),(6,'Jenny',90);
Using DENSE_RANK(), Calculate RANK of students:
;WITH cteStud AS
(
SELECT
StudName
,Totalmark
,DENSE_RANK() OVER (ORDER BY TotalMark DESC) AS StudRank
FROM tbl_Students
)
SELECT
StudName
,Totalmark
,StudRank
FROM cteStud
WHERE StudRank <= 3;
The Result:
studname | totalmark | studrank
----------+-----------+----------
Roy | 90 | 1
Jenny | 90 | 1
Anvesh | 88 | 2
Mahi | 88 | 2
Maria | 81 | 3
(5 rows)

TSQL invalid HAVING count

I am using SSMS 2008 and trying to use a HAVING statement. This should be a real simple query. However, I am only getting one record returned event though there are numerous duplicates.
Am I doing something wrong with the HAVING statement here? Or is there some other function that I could use instead?
select
address_desc,
people_id
from
dbo.address_view
where people_id is not NULL
group by people_id , address_desc
having count(*) > 1
sample data from address_view:
people_id address_desc
---------- ------------
Murfreesboro, TN 37130 F15D1135-9947-4F66-B778-00E43EC44B9E
11 Mohawk Rd., Burlington, MA 01803 C561918F-C2E9-4507-BD7C-00FB688D2D6E
Unknown, UN 00000 C561918F-C2E9-4507-BD7C-00FB688D2D6E
Jacksonville, NC 28546 FC7C78CD-8AEA-4C8E-B93D-010BF8E4176D
Memphis, TN 38133 8ED8C601-5D35-4EB7-9217-012905D6E9F1
44 Maverick St., Fitchburg, MA 8ED8C601-5D35-4EB7-9217-012905D6E9F1
The GROUP BY is going to lump your duplicates together into a single row.
I think instead, you want to find all people_id values with duplicate address_desc:
SELECT a.address_desc, a.people_id
FROM dbo.address_view a
INNER JOIN (SELECT address_desc
FROM dbo.address_view
GROUP BY address_desc
HAVING COUNT(*) > 1) t
ON a.address_desc = t.address_desc
using row_number and partition you can find the duplicate occurrences where row_num>1
select address_desc,
people_id,
row_num
from
(
select
address_desc,
people_id,
row_number() over (partition by address_desc order by address_desc) row_num
from
dbo.address_view
where people_id is not NULL
) x
where row_num>1

DB2 query group by id but with max of date and max of sequence

My table is like
ID FName LName Date(mm/dd/yy) Sequence Value
101 A B 1/10/2010 1 10
101 A B 1/10/2010 2 20
101 X Y 1/2/2010 1 15
101 Z X 1/3/2010 5 10
102 A B 1/10/2010 2 10
102 X Y 1/2/2010 1 15
102 Z X 1/3/2010 5 10
I need a query that should return 2 records
101 A B 1/10/2010 2 20
102 A B 1/10/2010 2 10
that is max of date and max of sequence group by id.
Could anyone assist on this.
-----------------------
-- get me my rows...
-----------------------
select * from myTable t
-----------------------
-- limiting them...
-----------------------
inner join
----------------------------------
-- ...by joining to a subselection
----------------------------------
(select m.id, m.date, max(m.sequence) as max_seq from myTable m inner join
----------------------------------------------------
-- first group on id and date to get max-date-per-id
----------------------------------------------------
(select id, max(date) as date from myTable group by id) y
on m.id = y.id and m.date = y.date
group by id) x
on t.id = x.id
and t.sequence = x.max_seq
Would be a simple solution, which does not take account of ties, nor of rows where sequence is NULL.
EDIT: I've added an extra group to first select max-date-per-id, and then join on this to get max-sequence-per-max-date-per-id before joining to the main table to get all columns.
I have considered your table name as employee..
check the below thing helped you.
select * from employee emp1
join (select Id, max(Date) as dat, max(sequence) as seq from employee group by id) emp2
on emp1.id = emp2.id and emp1.sequence = emp2.seq and emp1.date = emp2.dat
I'm a fan of using the WITH clause in SELECT statements to organize the different steps. I find that it makes the code easier to read.
WITH max_date(max_date)
AS (
SELECT MAX(Date)
FROM my_table
),
max_seq(max_seq)
AS (
SELECT MAX(Sequence)
FROM my_table
WHERE Date = (SELECT md.max_date FROM max_date md)
)
SELECT *
FROM my_table
WHERE Date = (SELECT md.max_date FROM max_date md)
AND Sequence = (SELECT ms.max_seq FROM max_seq ms);
You should be able to optimize this further as needed.