Using the provided table I would like to sample let's say 2 users per day so that users assigned to the two days are different. Of course the problem I have is more sophisticated, but this simple example gives the idea.
drop table if exists test;
create table test (
user_id int,
day_of_week int);
insert into test values (1, 1);
insert into test values (1, 2);
insert into test values (2, 1);
insert into test values (2, 2);
insert into test values (3, 1);
insert into test values (3, 2);
insert into test values (4, 1);
insert into test values (4, 2);
insert into test values (5, 1);
insert into test values (5, 2);
insert into test values (6, 1);
insert into test values (6, 2);
The expected results would look like this:
create table results (
user_id int,
day_of_week int);
insert into results values (1, 1);
insert into results values (2, 1);
insert into results values (3, 2);
insert into results values (6, 2);
You can use window functions. Here is an example . . . although the details do depend on your database (functions for random numbers vary by database):
select t.*
from (select t.*, row_number() over (partition by day_of_week order by random()) as seqnum
from test t
) t
where seqnum <= 2;
Related
I am trying to get distinct values with multi column select.
Sample table:
CREATE TABLE DUP_VALUES (ID NUMBER, NAME VARCHAR2(64));
INSERT INTO DUP_VALUES values (1, 'TEST1');
INSERT INTO DUP_VALUES values (2, 'TEST1');
INSERT INTO DUP_VALUES values (3, 'TEST2');
INSERT INTO DUP_VALUES values (4, 'TEST2');
INSERT INTO DUP_VALUES values (5, 'TEST1');
INSERT INTO DUP_VALUES values (6, 'TEST1');
INSERT INTO DUP_VALUES values (7, 'TEST1');
I want to get
ID NAME
1 TEST1
3 TEST2
I tried with SELECT DISTINCT ID, NAME FROM DUP_VALUES
But, I got all values, because ID is unique.
Use aggregation:
select min(id) as id, name
from dup_values
group by name;
I'm new to SQL Server and I searched for a solution to find, if a group is included in another group.
The query result should be grp_id 2 because 'A'+'B' is included in grp 3 and 5.
The result should be the grp_id of the the groups, that are included in other groups. With this result i´ll make an update of another table, joined with the grp_id.
The result should be:
+----+
| id |
+----+
| 2 |
+----+
I stuck in SQL because I do not find a solution to compare the groups. The idea was using bitwise comparison. But for that I had to add the value of each item in a field. I think there could be an easier way.
Thank you and best regards!
Eric
create table tmp_grpid (grp_id int);
create table tmp_grp (grp_id int, item_val nvarchar(10));
insert into tmp_grpid(grp_id) values (1);
insert into tmp_grpid(grp_id) values (2);
insert into tmp_grpid(grp_id) values (3);
insert into tmp_grpid(grp_id) values (4);
insert into tmp_grpid(grp_id) values (5);
--
insert into tmp_grp(grp_id, item_val) values (1, 'A');
insert into tmp_grp(grp_id, item_val) values (2, 'A');
insert into tmp_grp(grp_id, item_val) values (2, 'B');
insert into tmp_grp(grp_id, item_val) values (3, 'A');
insert into tmp_grp(grp_id, item_val) values (3, 'B');
insert into tmp_grp(grp_id, item_val) values (3, 'C');
insert into tmp_grp(grp_id, item_val) values (4, 'A');
insert into tmp_grp(grp_id, item_val) values (4, 'C');
insert into tmp_grp(grp_id, item_val) values (4, 'D');
insert into tmp_grp(grp_id, item_val) values (5, 'A');
insert into tmp_grp(grp_id, item_val) values (5, 'B');
insert into tmp_grp(grp_id, item_val) values (5, 'E');
Geez!
Technically speaking, group one is found in all other groups right? So, first a cross join to itself would be best with the condition that the values are the same AND that the groups are different, but before we do that we need to know how many items belong to each group so that's why we have the first select as a group that includes the count of elements per group, then join that with the cross join...Hope this helps.
select distinct dist_grpid
from
(select grp_id, count(*) cc from tmp_grp group by grp_id) g
inner join
(
select dist.grp_id dist_grpid, tmp_grp.grp_id, count(*) cc
from
tmp_grp dist
cross join tmp_grp
where
dist.item_val = tmp_grp.item_val and
dist.grp_id != tmp_grp.grp_id
group by
dist.grp_id,
tmp_grp.grp_id
) cj on g.grp_id = cj.dist_grpid and g.cc = cj.cc
I have a requirement with some business rules to implement on SQL (within a PL/SQL block): I need to evaluate such rules and according to the result perform the corresponding update, delete or insert into a target table.
My database model contains a "staging" and a "real" table. The real table stores records inserted in the past and the staging one contains "fresh" data coming from somewhere that needs to be merged into the real one.
Basically these are my business rules:
Delta between staging MINUS real --> Insert rows into the real
Delta between real MINUS staging--> Delete rows from the real
Rows which PK is the same but any other fields different: Update.
(Those "MINUS" will compare ALL the fields to get equality and distinguise the 3rd case)
I haven't figured out the way to accomplish such tasks without overlapping between rules by using a merge statement: Any suggestion for the merge structure? Is it possible to do it all together within the same merge?
Thank you!
If I understand you task correctly following code should do the job:
--drop table real;
--drop table stag;
create table real (
id NUMBER,
col1 NUMBER,
col2 VARCHAR(10)
);
create table stag (
id NUMBER,
col1 NUMBER,
col2 VARCHAR(10)
);
insert into real values (1, 1, 'a');
insert into real values (2, 2, 'b');
insert into real values (3, 3, 'c');
insert into real values (4, 4, 'd');
insert into real values (5, 5, 'e');
insert into real values (6, 6, 'f');
insert into real values (7, 6, 'g'); -- PK the same but at least one column different
insert into real values (8, 7, 'h'); -- PK the same but at least one column different
insert into real values (9, 9, 'i');
insert into real values (10, 10, 'j'); -- in real but not in stag
insert into stag values (1, 1, 'a');
insert into stag values (2, 2, 'b');
insert into stag values (3, 3, 'c');
insert into stag values (4, 4, 'd');
insert into stag values (5, 5, 'e');
insert into stag values (6, 6, 'f');
insert into stag values (7, 7, 'g'); -- PK the same but at least one column different
insert into stag values (8, 8, 'g'); -- PK the same but at least one column different
insert into stag values (9, 9, 'i');
insert into stag values (11, 11, 'k'); -- in stag but not in real
merge into real
using (WITH w_to_change AS (
select *
from (select stag.*, 'I' as action from stag
minus
select real.*, 'I' as action from real
)
union (select real.*, 'D' as action from real
minus
select stag.*, 'D' as action from stag
)
)
, w_group AS (
select id, max(action) as max_action
from w_to_change
group by id
)
select w_to_change.*
from w_to_change
join w_group
on w_to_change.id = w_group.id
and w_to_change.action = w_group.max_action
) tmp
on (real.id = tmp.id)
when matched then
update set real.col1 = tmp.col1, real.col2 = tmp.col2
delete where tmp.action = 'D'
when not matched then
insert (id, col1, col2) values (tmp.id, tmp.col1, tmp.col2);
I have the following 3 tables:
CREATE TABLE Tests (
Test_ID INT,
TestName VARCHAR(50));
INSERT INTO Tests VALUES (1, 'SQL Test');
INSERT INTO Tests VALUES (2, 'C# Test');
INSERT INTO Tests VALUES (3, 'Java Test');
CREATE TABLE Users (
[User_ID] INT,
UserName VARCHAR(50));
INSERT INTO Users VALUES (1, 'Joe');
INSERT INTO Users VALUES (2, 'Jack');
INSERT INTO Users VALUES (3, 'Jane');
CREATE TABLE UserTests (
ID INT,
[User_ID] INT,
Test_ID INT,
Completed INT);
INSERT INTO UserTests VALUES (1, 1, 1, 0);
INSERT INTO UserTests VALUES (2, 1, 2, 1);
INSERT INTO UserTests VALUES (3, 1, 3, 1);
INSERT INTO UserTests VALUES (4, 2, 1, 0);
INSERT INTO UserTests VALUES (5, 2, 2, 0);
INSERT INTO UserTests VALUES (6, 2, 3, 0);
INSERT INTO UserTests VALUES (7, 3, 1, 1);
INSERT INTO UserTests VALUES (8, 3, 2, 1);
INSERT INTO UserTests VALUES (9, 3, 3, 1);
I would like to create some rule/trigger so that when a new user gets added to the Users table, an entry for each Test and that user's Id will get added to the UserTests table.
Something like this if the new user ID is 5:
INSERT dbo.UserTest
(USER_ID, TEST_ID, Completed)
VALUES
(5, SELECT TEST_ID FROM Tests, 0)
That syntax is of course wrong but to give an idea of what I expect to happen.
So I expect that statement to add these values to the UserTests table:
User ID| Test ID| Completed
5 | 1 | 0
5 | 2 | 0
5 | 3 | 0
You can use after trigger for user table.
Create Trigger tr_user on Users
After Insert
AS Begin
INSERT UserTest(USER_ID, TEST_ID, Completed)
Select I.USER_ID, t.TEST_ID, 0
From Inserted I, Tests t
END
Here's a SQL Fiddle that finds missing records and inserts them.
SQL Fiddle
The SELECT:
select u.user_id, t.test_id, 0 as Completed
from users u
cross join tests t
where not exists (
select 1
from usertests ut
where ut.user_id = u.user_id and ut.test_id = t.test_id)
Adding insert into UserTests (User_Id, Test_Id, Completed) before the select will insert these records.
You can add a user id on to the where clause to do it for a single user if required. It is re-runnable so it won't re-insert test ids for a user that already has them, but will add new ones if new tests are introduced.
I have been trying to move away from using DECODE to pivot rows in Oracle 11g, where there is a handy PIVOT function. But I may have found a limitation:
I'm trying to return 2 columns for each value in the base table. Something like:
SELECT somethingId, splitId1, splitName1, splitId2, splitName2
FROM (SELECT somethingId, splitId
FROM SOMETHING JOIN SPLIT ON ... )
PIVOT ( MAX(splitId) FOR displayOrder IN (1 AS splitId1, 2 AS splitId2),
MAX(splitName) FOR displayOrder IN (1 AS splitName1, 2 as splitName2)
)
I can do this with DECODE, but I can't wrestle the syntax to let me do it with PIVOT. Is this even possible? Seems like it wouldn't be too hard for the function to handle.
Edit: is StackOverflow maybe not the right Overflow for SQL questions?
Edit: anyone out there?
From oracle-developer.net it would appear that it can be done like this:
SELECT somethingId, splitId1, splitName1, splitId2, splitName2
FROM (SELECT somethingId, splitId
FROM SOMETHING JOIN SPLIT ON ... )
PIVOT ( MAX(splitId) ,
MAX(splitName)
FOR displayOrder IN (1 AS splitName1, 2 as splitName2)
)
I'm not sure from what you provided what the data looks or what exactly you would like. Perhaps if you posted the decode version of the query that returns the data you are looking for and/or the definition for the source data, we could better answer your question. Something like this would be helpful:
create table something (somethingId Number(3), displayOrder Number(3)
, splitID Number(3));
insert into something values (1, 1, 10);
insert into something values (2, 1, 11);
insert into something values (3, 1, 12);
insert into something values (4, 1, 13);
insert into something values (5, 2, 14);
insert into something values (6, 2, 15);
insert into something values (7, 2, 16);
create table split (SplitID Number(3), SplitName Varchar2(30));
insert into split values (10, 'Bob');
insert into split values (11, 'Carrie');
insert into split values (12, 'Alice');
insert into split values (13, 'Timothy');
insert into split values (14, 'Sue');
insert into split values (15, 'Peter');
insert into split values (16, 'Adam');
SELECT *
FROM (
SELECT somethingID, displayOrder, so.SplitID, sp.splitname
FROM SOMETHING so JOIN SPLIT sp ON so.splitID = sp.SplitID
)
PIVOT ( MAX(splitId) id, MAX(splitName) name
FOR (displayOrder, displayOrder) IN ((1, 1) AS split, (2, 2) as splitname)
);