SQL Server Query group by analytic - sql

I have a table like this
Id scid name namesuffix nameId namesuffixid fullname
--------------------------------------------------------
1 1 a a 100 100 a
2 1 a b 100 101 ab
3 1 b c 101 102 abc
4 1 c d 102 103 abcd
5 2 e e 104 104 e
6 2 e f 104 105 ef
7 2 f g 105 106 efg
8 3 i i 107 107 i
9 3 i j 107 108 ij
10 3 j k 108 109 ijk
11 3 k l 109 110 ijkl
12 3 l m 110 111 ijklm
for each scid (group by scid)
select firstRow fullName
Last row fullName
Expected output
id scid fullname
-------------------
1 1 a
4 1 abcd
5 2 e
7 2 efg
8 3 i
12 3 ijklm
I tried first_value and last_value analytic functions, but the rows are repeating, didn't get expected result.
Any help appreciated.

Another option is to use ROW_NUMBER() and COUNT
select
id, scid, fullname
from (
select
*, row_number() over (partition by scid order by id) rn
, count(*) over (partition by scid) cnt
from
myTable
) t
where
rn = 1
or rn = cnt

You could use FIRST_VALUE and LAST_VALUE as you proposed:
SELECT scid,
FIRST_VALUE(id) OVER(PARTITION BY scid ORDER BY id
ROWS UNBOUNDED PRECEDING) AS id,
FIRST_VALUE(fullname) OVER(PARTITION BY scid ORDER BY id
ROWS UNBOUNDED PRECEDING) AS fullname
FROM tab_name
UNION
SELECT scid,
LAST_VALUE(id) OVER(PARTITION BY scid ORDER BY id
RANGE BETWEEN CURRENT ROW AND UNBOUNDED FOLLOWING) AS id,
LAST_VALUE(fullname) OVER(PARTITION BY scid ORDER BY id
RANGE BETWEEN CURRENT ROW AND UNBOUNDED FOLLOWING) AS fullname
FROM tab_name
ORDER BY scid, id;
Demo

There are other ways to do this without window funtions:
select t.*
from t join
(select min(id) as min_id, max(id) as max_id
from t
group by sc_id
) tt
on t.id in (min_id, max_id);
I only suggest this because there are many ways to do what you want. If performance is an issue, you may want to experiment with different methods.

Related

How to find most frequent code(varchar) from a table

I would like to find most frequent code within CodeID which is in same code_group from a table.
For example, from original table
ID CodeID Name Code Code_group
1 1 A 101 0
2 1 A 102 0
3 1 B 102 0
4 2 C 201 0
5 2 C 201 0
6 2 D 202 0
7 2 E 202 0
8 3 F 101 1
9 3 G 103 1
10 3 G 104 1
11 3 G 104 1
I want output like the below.
ID CodeID Name Code Code_group Selected_code
1 1 A 101 0 102
2 1 A 102 0 102
3 1 B 102 0 102
4 2 C 201 0 NULL
5 2 C 201 0 NULL
6 2 D 202 0 NULL
7 2 E 202 0 NULL
8 3 F 101 1 104
9 3 G 103 1 104
10 3 H 104 1 104
11 3 H 104 1 104
Even though code of 8th ID is same in CodeID: 1,it is not in the same Code_group.
So For CodeID: 1, Selected_code would be 102.
it must be counted within exactly same Code_group.
=======================================
I have tried it like the below. I should not use ID for this one.
From TableA
with m as
(
select
CodeID,
Name,
Code,
Code_group,
cnt,
Selected_code = ROW_NUMBER() over (partition by Code_group order by cnt desc)
from( select CodeID, Name, Code,Code_group
,count(*) over (partition by Code,CodeID) as cnt from tableA
group by CodeID, Name, Code, Code_group,
) as t
group by CodeID,
Name,
Code,
Code_group, cnt
)
select a.CodeID,
a.Name,
a.Code,
a.Code_group, b.Code as Selected_code, cnt
from(select
CodeID,
Name,
Code,
Code_group,Selected_code,
cnt
from m) as a left outer join
(select CodeID,
Name,
Code,
Code_group,Selected_code,
cnt
from m where selected_Code=1) as b on a.CodeID = b.CodeID and a.Code_Group = b.Code_Group
order by a.CodeID, a.Code_Group
The problem of this is
With statment makes my table distinct. It shows only one row if there is exactly same data such as ID 1,2.
Also, I cannot make NULL if there is exactly same frequencies.
What should I add to get my desired output?
Or is there any better approach for this?
CTE cte find the highest frequency code by Code_group and CodeID using dense_rank()
CTE selected check for any Code with same frequency and exclude them.
Final query just select from the original table and LEFT JOIN the selected
with
cte as
(
select Code_group, CodeID, Code
from
(
select Code_group, CodeID, Code,
r = dense_rank() over (partition by Code_Group, CodeID
order by count(*) desc)
from tableA
group by Code_group, CodeID, Code
) c
where c.r = 1
),
selected as
(
select Code_group, CodeID, Code
from
(
select Code_group, CodeID, Code,
cnt = count(*) over (partition by Code_group, CodeID)
from cte
) s
where s.cnt = 1
)
select a.*,
Selected_Code = s.Code
from tableA a
left join selected s on a.Code_Group = s.Code_Group
and a.CodeID = s.CodeID;
db<>fiddle demo

DB2 Toad SQL - Group by Certain Columns using Max Command

I am having some trouble with the below query. I do understand I need to group by ID and Category, but I only want to group by ID while keeping the rest of the columns based on Rank being max. Is there a way to only group by certain columns?
select ID, Category, max(rank)
from schema.table1
group by ID
Input:
ID Category Rank
111 3 4
111 1 5
123 5 3
124 7 2
Current Output
ID Category Rank
111 3 4
111 9 1
123 5 3
124 7 2
Desired Output
ID Category Rank
111 1 5
123 5 3
124 7 2
You can use:
select *
from table1
where (id, rank) in (select id, max(rank) from table1 group by id)
Result:
ID CATEGORY RANK
---- --------- ----
111 1 5
123 5 3
124 7 2
Or you can use the ROW_NUMBER() window function. For example:
select *
from (
select *,
row_number() over(partition by id order by rank desc) as rn
from table1
) x
where rn = 1
See running example at db<>fiddle.
You can try using - row_number()
select * from
(
select ID, Category,rank, row_number() over(partition by id order by rank desc) as rn
from schema.table1
)A where rn=1

Find first N rows that have unique value with mod(id, N)

For example, N is 10 and a table looks like
id
1
2
3
4
5
6
7
10
11
12
13
108
109
111
112
113
Need to find first N rows that have unique value with mod(id, N).
Expected result is
mod10
1
2
3
4
5
6
7
10
108
109
I've tried something like
select *
from
(
select id, id % 10 as seq_id
from accounts order by id
) as s1
group by s1.seq_id limit 10;`
but not working.
You can use window function here -
SELECT id
FROM (SELECT id, ROW_NUMBER() OVER(PARTITION BY id % 10 ORDER BY id) RN
FROM table_name
) X
WHERE RN = 1
ORDER BY id
You can try the below one - using row_number()
DEMO
with cte as
(
select id, row_number() over(partition by seq_id order by id) as rn
from
(
select id, id % 10 as seq_id from tablename
)A
)
select id from cte where rn=1 order by id
OUTPUT:
id
1
2
3
4
5
6
7
10
108
109

How to use this in sql -- > max(sum (paid * quantity )) to solve a query

How to get the max value order of each customer ?
select num, max(sum(paid*quantity))
from orders join
pizza
using (order#)
group by customer#;
table
num orderN price
-------- --- -------
1 109 30
1 118 25
3 101 30
3 115 27
4 107 23
5 100 17
5 129 16
output req-
num Pnum price
-------- --- -------
1 109 30
3 101 30
4 107 23
5 100 17
You want to select the record having the highest price in each group of nums.
If your RDBMS supports window functions, that's straight forward with ROW_NUMBER() :
SELECT num, pnum, price
FROM (
SELECT t.*, ROW_NUMBER OVER(PARTITION BY num ORDER BY price DESC) rn
FROM mytable t
) x
WHERE rn = 1
Else, you can take the following approach, that uses a NOT EXISTS condition with a correlated subquery to ensure that the record being joined in the one with the highest price for the current num :
SELECT num, pnum, price
FROM mytable t
WHERE NOT EXISTS (
SELECT 1 FROM mytable t1 WHERE t1.num = t.num AND t1.price > t.price
)

Select and aggregate last records base on order

I have different versions of the charges in a table. I want to grab and sum the last charge grouped by Type.
So I want to add 9.87, 9.63, 1.65.
I want the Parent ID , sum(9.87 + 9.63 + 1.65) as the results of this query.
We use MSSQL
ID ORDER CHARGES TYPE PARENT ID
1 1 6.45 1 1
2 2 1.25 1 1
3 3 9.87 1 1
4 1 6.54 2 1
5 2 5.64 2 1
6 3 0.84 2 1
7 4 9.63 2 1
8 1 7.33 3 1
9 2 5.65 3 1
10 3 8.65 3 1
11 4 5.14 3 1
12 5 1.65 3 1
WITH recordsList
AS
(
SELECT Type, Charges,
ROW_NUMBER() OVER (PArtition BY TYPE
ORDER BY [ORDER] DESC) rn
FROM tableName
)
SELECT SUM(Charges) totalCharge
FROM recordsLIst
WHERE rn = 1
SQLFiddle Demo
Use row_number() to identify the rows to be summed, and then sum them:
select SUM(charges)
from (select t.*,
ROW_NUMBER() over (PARTITION by type order by id desc) as seqnum
from t
) t
where seqnum = 1
Alternatively you could use a window aggregate MAX():
SELECT SUM(Charges)
FROM (
SELECT
[ORDER],
Charges,
MaxOrder = MAX([ORDER]) OVER (PARTITION BY [TYPE])
FROM atable
) s
WHERE [ORDER] = MaxOrder
;
SELECT t.PARENT_ID, SUM(t.CHARGES)
FROM dbo.test73 t
WHERE EXISTS (
SELECT 1
FROM dbo.test73
WHERE [TYPE] = t.[TYPE]
HAVING MAX([ORDER]) = t.[ORDER]
)
GROUP BY t.PARENT_ID
Demo on SQLFiddle