SQL Server being used to run this
I'm trying to pull information that could be variable.. Could you please check it and find the error?

  id         = t.id
  name       = c.name
  fav_flavor = ISNULL((SELECT value FROM ce WHERE key='flavor'), '')
  fav_color  = ISNULL((SELECT value FROM ce WHERE key='color'), '')
  Transaction t
  LEFT JOIN Customer ON t.customer_id=c.id
  LEFT JOIN CustomerExtra ce ON ce.id=c.id

Of course this is a siplified query to be used in this example, but it has the scense.

So, basically, I'm not able to use that ce table referenced in the JOIN statement to get random values (in this case 'flavor' and 'color') in the main query.
Is there a way to access those values without adding the condition in the last JOIN and having to duplicate that line as many times as I need values from it? or... is it the right way?

  id         = t.id,
  name       = c.name,
  fav_flavor = ce1.value,
  fav_color  = ce2.value,
  cust_other = ce3.value
  Transaction t
  LEFT JOIN CustomerExtra ce1 ON ce1.id=c.id AND ce1.key='flavor'
  LEFT JOIN CustomerExtra ce2 ON ce2.id=c.id AND ce2.key='color'
  LEFT JOIN CustomerExtra ce3 ON ce3.id=c.id AND ce3.key='other'
  • 2
    Your CustomerExtra table sounds like it's designed with the EAV anti-pattern which is part of your problem. Generally, EAV should be avoided. In a more normalized design, you'd be able to solve this problem more easily.
    – J.D.
    Jul 21 at 21:30

1 Answer 1


Use Outer apply:

  Color = max(case when key = 'color' then value else '' end)
,flavor = max(case when key = 'flavor' then value else '' end)
CustomerExtra ce
Ce.id = c.id
) CE

de here

Because for each customer errow you have muktiple extra key-pair values, using a aggregate function allows you just pivot this vakues to columns.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.