I'm updating 3 lookup tables' association tables like so:
insert into LookUpTable1_AssociationTable1 ([LookupId],[AssociationId])
select Id as LookupId, 4 as AssociationId from LookupTable1
insert into LookUpTable2_AssociationTable2 ([LookupId],[AssociationId])
select Id as LookupId, 4 as AssociationId from LookupTable2
I have 4 records in my Association master table. So I can run the above replacing the hard-coded '4' with each id in my association master table, but can i do it as a set somehow? Just run a "set based" (not procedural) sql that takes all 4 records in association table and performs the statements above automatically
Not sure if i understand your intent correctly. Check whether below query suits your needs.
insert into LookUpTable1_AssociationTable1 ([LookupId],[AssociationId])
select
l.Id as LookupId
,a.Id as AssociationId
from
LookupTable1 l
cross join Association a
Related
Suppose,
I have a table tblClasses and tblStudents
Now each class have multiple student.
tblClass
ID Name
1 ClassA
2 ClassB
3 ClassC
tblStudents
ID Name Class_ID
1. john 1
2. Mathew 1
3. Imran 2
4. Jenny 3
now, I have another server having exact same db and tables and I am copying data from server 1 to server 2 from same tables using the Select and Insert e.g.
insert into server2.dbo.tblClass (Name)
select Name from server1.dbo.tblClass
and for tblStudents
insert into server2.dbo.tblStudents (Name, Class_ID)
select Name, Class_ID from server1.dbo.tblStudents
now this is ok but the real problem is that in server2 after copying the data, how to populate the tblStudents fk Class_ID with the actual IDs of tblClass which are generated after inserting the data into tblStudents in server2 since PKs are Identity and autoincremented and cannot change the design.
What to do in this case?
In simple words, when a parent and child data are copied then in the child table the foreign key field needs to be populated with the actual IDs of the parent not the one from where it is copied which would be obviously different.
I am not allowed to change the table design or properties and have to do it using the queries.
Any suggestions?
The way is to create a ClassId mapping table on class records insertion and use this mapping table to translate OldClassId to NewClassId for the new Student table:
declare #ClassIds table (OldClassId int, NewClassId int);
merge into newDB.dbo.tblClasses as target
using
(
select
Id = Id * (-1),
[Name]
from
oldDB.dbo.tblClasses
)
as source on source.Id = target.Id
when not matched by target then
insert ([Name])
values (source.[Name])
output source.Id * (-1), inserted.Id -- ← the trick is here
into #ClassIds (OldClassId, NewClassId);
insert into newDB.dbo.tblStudents
select
s.Id,
s.[Name],
ClassId = ids.NewClassId
from
oldDB.dbo.tblStudents s
inner join #ClassIds ids on ids.OldClassId = s.ClassId;
The major trick is that the MERGE statement may work with not not only inserted and deleted columns (as INSERT, UPDATE, DELETE statements do) but with the source columns as well.
There are two ways to go that I can think:
The first is to use SET IDENTITY_INSERT tblClass ON. This is not a design change, so you should be able to do it. After that, you can insert your own values in tblClass.ID (although you will need the select list parenthesis):
insert tblClass(ID,Name) values (1, 'ClassA')....
Alternatively, you could make a query that connects student ids to class names, and then use that back to create the respective connections:
-- export/save this in temp table
select s.Name,c.Name as className
into #a
from tblStudents s
left join tblClass c on s.Class_ID=c.ID
--now use this to fill db2 tblStudents
insert tblStudents(Name,Class_ID)
select #a.Name,c.ID
from
#a
inner join tblClass c on #a.className=c.Name
I am new to SQL and I would like to get some insights for my problem
I am using the following query,
select id,
pid
from assoc
where id in (100422, 100414, 100421, 100419, 100423)
All these id need not have pid, some doesn't and some has pid. Currently it skips the records which doesn't have pid.
I would like a way which would show the results as below.
pid id
-----------
703 100422
313 100414
465 100421
null 100419
null 100423
Any help would be greatly appreciated. Thanks!
Oh, I think I've got the idea: you have to enumerate all the ids and corresponding pids. If there's no corresponding pid, put null (kind of outer join). If it's your case, then Oracle solution can be:
with
-- dummy: required ids
dummy as (
select 100422 as id from dual
union all select 100414 as id from dual
union all select 100421 as id from dual
union all select 100419 as id from dual
union all select 100423 as id from dual),
-- main: actual data we have
main as (
select id,
pid
from assoc
-- you may put "id in (select d.id from dummy d)"
where id in (100422, 100414, 100421, 100419, 100423))
-- we want to print out either existing main.pid or null
select main.pid as pid,
dummy.id as id
from dummy left join main on dummy.id = main.id
id is obtained from other table and assoc only has pid associated with id.
The assoc table seems to be the association table used to implement a many-to-many relationship between two entities in a relational database.
It contains entries only for the entities from one table that are in relationship with entities from the other table. It doesn't contain information about the entities that are not in a relationship and some of the results you want to get come from entities that are not in a relationship.
The solution for your problem is to RIGHT JOIN the table where the column id comes from and put the WHERE condition against the values retrieved from the original table (because it contains the rows you need). The RIGHT JOIN ensures all the matching rows from the right side table are included in the result set, even when they do not have matching rows in the left side table.
Assuming the table where the id column comes from is named table1, the query you need is:
SELECT assoc.id, assoc.pid
FROM assoc
RIGHT JOIN table1 ON assoc.id = table1.id
WHERE table1.id IN (100422, 100414, 100421, 100419, 100423)
I have two tables bottle and case that are filled out. They both have a column labeled case_id however in bottle table all the values are 0 while in case table they have the correct id.
How can I update the first table bottle values 0 with the new values from other table case? I believe I will need to use an UPDATE or INSERT or INNER JOIN.
If you give more detail for table struct will be better
UPDATE b SET b.case_id=c.case_id
FROM bottle AS b INNER JOIN [Case] AS c ON b.some_coloumn=c.some_cloumn
Say I have two customer tables which have pretty much the same columns. One of them is a temporal one which periodically updates the other one. That is, updates and additions to the records are done to the temporal table only. Say they're names are CUSTOMER, and CUSTOMER_TEMP.
The tables got information like ID, NAME, LAST_NAME, and ADDRESS.
This temporal table has three extra fields, TEMP_ID, RECORD_TYPE, and DATE. The record type is used to record whether there was an addition or an update. So the thing is I need to select the latest record from both tables. That involves several cases
Main table has record, but temporal doesn't -> Select main table record.
Main table has no record, but temporal does -> Select latest temporal table record.
Main table has record and temporal has an add record -> Select temporal table record.
Main table has record and temporal table has update record -> Select temporal table record.
Main table has record and temporal table has add and update record. -> Select temporal table update record.
Main table has record and temporal table has various update records. -> Select latest temporal table update record.
Main table has record and temporal table has add record and various update records. -> Select latest temporal table update record.
Now, I don't know whether this is a good flow or not. I was just told to make the query, so I don't have access to the DB, I believe I could make suggestions though. The thing is My SLQ knowledge is not enough to build this query. I know there's an INNER_JOIN involved, as well as a filter by date, and probably and EXIST, to check whether the record exist or not in the CUSTOMER_TEMP table. But I don't quite know how to build it. I'm working on .Net And SQLServer. Any help on it is quite appreciated.
select m.*, 0 as [rn]
from main m
where not exists (select 1 from temp where temp.id = m.id)
union
select tt.*
from ( select temp.*
, row_number() over (partition by id order by RECORD_TYPE desc, date desc) as rn
from temp
-- join main
-- on temp.ID = main.ID
) tt
where tt.rn = 1
if update does not sort last then need to do a trick like in the answer from Tom H
;WITH CTE_Latest_Temporal AS
(
SELECT
id,
name,
..., -- Put the rest of your columns here
ROW_NUMBER OVER (PARTITION BY id
ORDER BY
CASE record_type
WHEN 'Update' THEN 0
ELSE 1
END, date DESC) AS row_num
FROM
Customer_Temp
)
SELECT
M.id,
CASE WHEN T.id IS NOT NULL THEN T.name ELSE M.name END AS name,
... -- Similar CASE statements for the rest of your columns
FROM
Customer M
LEFT OUTER JOIN CTE_Latest_Temporal T ON
T.id = M.id AND
T.row_num = 1
The CASE statements can be replaced by a simple COALESCE(T.column, M.column) for any columns that cannot be NULL. I had to use the CASE to cover situations where the row might exist in the temp table but the column might be NULL in the temp table, but have a value in the main table.
The outline of the tables in question are as follows:
I have a table, lets call it join, that has two columns, both foreign keys to other tables. Let's call the two columns userid and buildingid so join looks like
+--------------+
| join |
|--------------|
|userid |
|buildingid |
+--------------+
I basically need to insert a bunch of rows into this table. Each user will be assigned to multiple buildings by having multiple entries in this table. So user 13 might be assigned to buildings 1, 2, and 3 by the following
13 1
13 2
13 3
I'm trying to figure out how to do this in a query if the building numbers are constant, that is, I'm assigning a group of people to the same buildings. Basically, (this is wrong) I want to do
insert into join (userid, buildingid) values ((select userid from users), 1)
Does that make sense? I've also tried using
select 1
The error I'm running into is that the subquery returns more than one result. I also attempted to create a join, basically with a static select query that was also unsuccessful.
Any thoughts?
Thanks,
Chris
Almost! When you want to insert to values of a query, don't try to put them in the values clause. insert can take a select as an argument for the values!
insert into join (userid, buildingid)
select userid, 1 from users
Also, in the spirit of learning more, you can create a table that doesn't exist by using the following syntax:
select userid, 1 as buildingid
into join
from users
That only works if the table doesn't exist, though, but it's a quick and dirty way to create table copies!