SQL Update Multiple Fields FROM via a SELECT Statement

Question

This works, but i would like to remove the redundancy. Is there a way to merge the update with a single select statement so i don\'t have to use vars?

    DE         


        

Answer 1

I would write it this way

UPDATE s
SET    OrgAddress1 = bd.OrgAddress1,    OrgAddress2 = bd.OrgAddress2,    
     ...    DestZip = bd.DestZip
--select s.OrgAddress1, bd.OrgAddress1, s.OrgAddress2, bd.OrgAddress2, etc 
FROM    Shipment s
JOIN ProfilerTest.dbo.BookingDetails bd on  bd.MyID =s.MyID2
WHERE    bd.MyID = @MyId 

This way the join is explicit as implicit joins are a bad thing IMHO. You can run the commented out select (usually I specify the fields I'm updating old and new values next to each other) to make sure that what I am going to update is exactly what I meant to update.

Answer 2

you can use update from...

something like:

update shipment set.... from shipment inner join ProfilerTest.dbo.BookingDetails on ...

Answer 3

You should be able to do something along the lines of the following

UPDATE s
SET
    OrgAddress1 = bd.OrgAddress1,
    OrgAddress2 = bd.OrgAddress2,
    ...
    DestZip = bd.DestZip
FROM
    Shipment s, ProfilerTest.dbo.BookingDetails bd
WHERE
    bd.MyID = @MyId AND s.MyID2 = @MyID2

FROM statement can be made more optimial (using more specific joins), but the above should do the trick. Also, a nice side benefit to writing it this way, to see a preview of the UPDATE change UPDATE s SET to read SELECT! You will then see that data as it would appear if the update had taken place.

Answer 4

I just had to solve a similar problem where I added a Sequence number (so that items as grouped by a parent ID, have a Sequence that I can order by (and presumably the user can change the sequence number to change the ordering).

In my case, it's insurance for a Patient, and the user gets to set the order they are assigned, so just going by the primary key isn't useful for long-term, but is useful for setting a default.

The problem with all the other solutions is that certain aggregate functions aren't allowed outside of a SELECT

This SELECT gets you the new Sequence number:

select PatientID,
       PatientInsuranceID, 
       Sequence, 
       Row_Number() over(partition by PatientID order by PatientInsuranceID) as RowNum
from PatientInsurance
order by PatientID, PatientInsuranceID

This update command, would be simple, but isn't allowed:

update PatientInsurance
set Sequence = Row_Number() over(partition by PatientID order by PatientInsuranceID)

The solution that worked (I just did it), and is similar to eKek0's solution:

UPDATE PatientInsurance
SET  PatientInsurance.Sequence = q.RowNum
FROM (select PatientInsuranceID, 
             Row_Number() over(partition by PatientID order by PatientInsuranceID) as RowNum
      from PatientInsurance
     ) as q 
WHERE PatientInsurance.PatientInsuranceID=q.PatientInsuranceID 

this lets me select the ID I need to match things up to, and the value I need to set for that ID. Other solutions would have been fine IF I wasn't using Row_Number() which won't work outside of a SELECT.

Given that this is a 1 time operation, it's coding is still simple, and run-speed is fast enough for 4000+ rows

Answer 5

You can use:

UPDATE s SET
  s.Field1 = q.Field1,
  s.Field2 = q.Field2,
  (list of fields...)
FROM (
  SELECT Field1, Field2, (list of fields...)
  FROM ProfilerTest.dbo.BookingDetails 
  WHERE MyID=@MyID
) q
WHERE s.MyID2=@ MyID2

Answer 6

Something like this should work (can't test it right now - from memory):

UPDATE SHIPMENT
SET
  OrgAddress1     = BD.OrgAddress1,
  OrgAddress2     = BD.OrgAddress2,
  OrgCity         = BD.OrgCity,
  OrgState        = BD.OrgState,
  OrgZip          = BD.OrgZip,
  DestAddress1    = BD.DestAddress1,
  DestAddress2    = BD.DestAddress2,
  DestCity        = BD.DestCity,
  DestState       = BD.DestState,
  DestZip         = BD.DestZip
FROM
   BookingDetails BD
WHERE 
   SHIPMENT.MyID2 = @MyID2
   AND
   BD.MyID = @MyID

Does that help?