Question

I have the following sample table structures for a parent / child relationship.

parent table

id    LongText
--------------
10    
20    

child table

id    char     value
--------------------
10    COLOR    RED
10    HEIGHT   1FT
20    COLOR    BLUE
20    WIDTH    2FT

I have a requirement that specifies that certain fields from the child table rows need to be concatenated together and placed in the parent table. I would like to complete this in a single SQL query if possible. The update statement that I have written is as follows.

UPDATE
  parent
SET
  LongText = COALESCE(LongText, N'')
              + child.char + N': ' + child.val + ','
FROM
  parent INNER JOIN child
    ON
      parent.id = child.id

But I only get the following results.

id    LongText
------------------
10    COLOR: RED,
20    COLOR: BLUE,

I would expect (or I should say I WANT) to get this.

id    LongText
------------------
10    COLOR: RED,HEIGHT: 1FT
20    COLOR: BLUE,WIDTH: 2FT

Is this possible? Any suggestions as to how I can do this? Any help is appreciated!

Here is a SQLfiddle for reference.

Answer 1

update parent
set longtext =
    stuff((
    select ',' + c.char + ': ' + c.val
    from child c
    where c.uid = parent.uid
    for xml path(''), type).value('.','nvarchar(max)'),1,1,'');

I have updated your SQLFiddle with the solution.

1. STUFF() function to remove the leading ',' (comma) from the first characteristic.

2. FOR XML to create an XML document from a query result. This is a well known trick with SQL Server – because the column is not named, there is no element produced and only the raw text (of each row) is output, mashed together into a single row.

Very few articles on the internet attempt to explain it in detail, since the code is pretty much left as an explanation in and of itself.