May 1, 2018 at 10:13 pm
I have three tables with user data in them. The formatting might be off but hopefully below should give you the relation.
Users Table
UserID | FirstName | LastName | MiddleName
1 | Some | Dude | Big
2 | Another | Dude | Little
UDF Table
UserID | FieldNo | Value
1 | 1 | 123456
1 | 2 | Text1
1 | 3 | 01/01/2018
1 | 4 | Text2
1 | 5 | Text3
2 | 1 | 234567
2 | 2 | Value1
2 | 3 | 12/12/2012
2 | 4 | Value2
2 | 5 | Value3
UserImport Table (Created for this sole purpose)
UserID | FirstName | LastName | MiddleName | UDF1 | UDF2 | UDF3 | UDF4 | UDF5
1 | Some | Dude | Big | 123456 | Text1 | 01/01/2018 | Text2 | Text3
2 | Another | Dude | Little | 234567 | Value1 | 12/12/2012 | Value2 | Value3
Iām using SSIS to parse information coming into these tables. An example would be that UserId 1 will have a change to the UDF2 and UDF3 fields. I bring the information into the UserImport table and was planning on doing an update to the UDF table from there. I cannot seem to get my head wrapped around updating the un-pivoted table. How do I update the row instead of the column? Below is as close as I could get and I think what I have is way off. Any help would be greatly appreciated.
Update UDF
SET
Value = CASE WHEN FieldNo = 2 Then (Select UDF3 From UsersImport )
WHEN FieldNo = 3 Then (Select UDF3 From UsersImport)
END
Where UDF.UserID = (Select UserID FROM UserImport)
May 1, 2018 at 11:25 pm
create table users
(
userid int,
firstname varchar(25),
lastname varchar(25),
middlename varchar(25)
);
insert into users values (1,'some','dude','big');
create table udf
(
userid int,
fieldno int,
[value] varchar(25)
);
insert into udf values (1,1,123456);
insert into udf values (1,2,'text1'); create table userimport
(
userid int,
firstname varchar(25),
lastname varchar(25),
middlename varchar(25),
udf1 varchar(25),
udf2 varchar(25),
udf3 varchar(25),
udf4 varchar(25),
udf5 varchar(25),
)
Kindly confirm whether it is working or not as i was doing it in SQL fiddle
update userimport
set
userid =a.userid,
firstname=a.firstname,
lastname=a.lastname,
middlename=a.middlename,
udf1= case when b.fieldno=1 then [value] end ,
udf2= case when b.fieldno=2 then [value] end
from
users a inner join udf b on a.userid=b.userid
Saravanan
May 2, 2018 at 8:20 am
I believe your update query works but I didn't explain well enough that the I am trying to update the UDF table and not the UserImport table. To be specific I am trying to update the Value column. Below is my latest failed attempt. I'm starting to have my doubts about using CASE due to need for different information in the same column. Thanks for any additional insight.
Update F
Set
[Value] = CASE WHEN F.UserID = I.UserID and F.FieldNo = 1 then I.UDF1
WHEN F.UserID = I.UserID and F.FieldNo = 2 then I.UDF2
WHEN F.UserID = I.UserID and F.FieldNo = 3 then I.UDF3
END
FROM UDF as F INNER JOIN Users as U
ON U.UserId = F.UserID INNER JOIN UserImport as I
ON I.UserID = F.UserID and I.UDF1 = F.[Value] WHERE F.FieldNo = 1
May 2, 2018 at 11:26 am
what i have done is to convert the records in userimport from columns as rows first.
ie (udf1, udf2, ...) is converted as
userid,fieldno,value
1 , 1 , New1
1 , 2 , New2
1 , 3 , New3
1 , 4 , New4
1 , 5 , New5
So the result of that step would be how we want the final output to look like.
This is followed by a update using join with the result set in step 1 with the udf table on the keys userid,fieldno.
(You could also accomplish the same using MERGE statement as well)
create table users
(
userid int,
firstname varchar(25),
lastname varchar(25),
middlename varchar(25)
);
insert into users values (1,'some','dude','big');
create table udf(userid int,fieldno int, value varchar(25));
insert into udf values (1,1,'123456');
insert into udf values (1,2,'text1');
create table userimport
(
userid int
,firstname varchar(25)
,lastname varchar(25)
,middlename varchar(25)
,udf1 varchar(25)
,udf2 varchar(25)
,udf3 varchar(25)
,udf4 varchar(25)
,udf5 varchar(25)
);
insert into userimport values(1,'some','dude','big','New1','New2','New3','New4','New5');
update a2
set a2.value=b2.value1
from udf a2
join (select a.userid
,b.rnk as fieldno
,case when b.rnk=1 then a.udf1
when b.rnk=2 then a.udf2
when b.rnk=3 then a.udf3
when b.rnk=4 then a.udf4
when b.rnk=5 then a.udf5
end as Value1
from userimport a
join (select top 5 row_number() over(order by (select null)) as rnk
from information_schema.tables
)b
on 1=1
)b2
on a2.userid=b2.userid
and a2.fieldno=b2.fieldno;
select *
from udf;
May 3, 2018 at 8:48 am
You are a genius, George. That seems to work very well. Thanks for your help.
May 3, 2018 at 10:30 am
Cool. Glad i could help š
May 3, 2018 at 11:19 am
This should be enough to UPDATE the udf table:
UPDATE udf
SET Value = CASE udf.fieldno WHEN 1 THEN udf1 WHEN 2 THEN udf2 WHEN 3 THEN udf3
WHEN 4 THEN udf4 WHEN 5 THEN udf5 END
FROM udf
INNER JOIN UserImport UI ON UI.userid = udf.userid
Full example:
use tempdb;
create table users
(
userid int,
firstname varchar(25),
lastname varchar(25),
middlename varchar(25)
);
insert into users values (1,'some','dude','big');
create table udf(userid int,fieldno int, value varchar(25));
insert into udf values (1,1,'123456');
insert into udf values (1,2,'text1');
insert into udf values (1,3,'01/01/2018');
insert into udf values (1,4,'Text2');
insert into udf values (1,5,'Text3');
create table userimport
(
userid int
,firstname varchar(25)
,lastname varchar(25)
,middlename varchar(25)
,udf1 varchar(25)
,udf2 varchar(25)
,udf3 varchar(25)
,udf4 varchar(25)
,udf5 varchar(25)
);
insert into userimport values(1,'some','dude','big','New1','New2','New3','New4','New5');
SELECT 'Before', * FROM udf
UPDATE udf
SET Value = CASE udf.fieldno WHEN 1 THEN udf1 WHEN 2 THEN udf2 WHEN 3 THEN udf3
WHEN 4 THEN udf4 WHEN 5 THEN udf5 END
FROM udf
INNER JOIN UserImport UI ON UI.userid = udf.userid
SELECT 'After', * FROM udf
drop table users
drop table udf
drop table userimport
SQL DBA,SQL Server MVP(07, 08, 09) A socialist is someone who will give you the shirt off *someone else's* back.
May 3, 2018 at 10:50 pm
Nice one , that is a better and simple solution than mine.
Viewing 8 posts - 1 through 7 (of 7 total)
You must be logged in to reply to this topic. Login to reply