SQL UPDATE row Number
Solution 1
You have to rewrite it using UPDATE FROM, the syntax is just a bit bulky:
UPDATE serviceClustersNew
FROM
(
SELECT text_id,
(SELECT MAX(ID) FROM serviceClusters) +
ROW_NUMBER() OVER (ORDER BY Text_ID) AS newID
FROM serviceClustersNew
) AS src
SET ID = newID
WHERE serviceClustersNew.Text_ID = src.Text_ID
Solution 2
You are not dealing with a lot of data, so a correlated subquery can serve the same purpose:
UPDATE serviceClustersNew
SET ID = (select max(ID) from serviceClustersNew) +
(select count(*)
from serviceClustersNew scn2
where scn2.Text_Id <= serviceClustersNew.TextId
)
This assumes that the text_id
is unique along the rows.
el smu
Updated on June 04, 2022Comments
-
el smu almost 2 years
I have a table serviceClusters with a column identity(1590 values). Then I have another table serviceClustersNew with the columns ID, text and comment. In this table, I have some values for text and comment, the ID is always 1. Here an example for the table:
[1, dummy1, hello1;
1, dummy2, hello2;
1, dummy3, hello3;
etc.]
WhaI want now for the values in the column ID is the continuing index of the table serviceClusters plus the current Row number: In our case, this would be 1591, 1592 and 1593.
I tried to solve the problem like this: First I updated the column ID with the maximum value, then I tryed to add the row number, but this doesnt work:
-- Update ID to the maximum value 1590 UPDATE serviceClustersNew SET ID = (SELECT MAX(ID) FROM serviceClusters); -- This command returns the correct values 1591, 1592 and 1593 SELECT ID+ROW_NUMBER() OVER (ORDER BY Text_ID) AS RowNumber FROM serviceClustersNew -- But I'm not able to update the table with this command UPDATE serviceClustersNew SET ID = (SELECT ID+ROW_NUMBER() OVER (ORDER BY Text_ID) AS RowNumber FROM serviceClustersNew)
By sending the last command, I get the error "Syntax error: Ordered Analytical Functions are not allowed in subqueries.". Do you have any suggestions, how I could solve the problem? I know I could do it with a volatile table or by adding a column, but is there a way without creating a new table / altering the current table?