Question

I have the following query:

SELECT
  tableOneId
  SUM(a+b+c) AS tableOneData,
  MIN(d) AS tableTwoData,
FROM
  tableTwo JOIN tableOne ON tableOneId = tableTwoId
GROUP BY
  tableOneId

All of the mentioned columns are declared as numeric(30,6) NOT NULL.

In tableOne, I have entries whose sum (columns a, b, c) should be equivalent to column d in Table Two.

A simple example of this:

Table One (id here should read tableOneId to match above query)
  id=1, a=1, b=0, c=0
  id=1, a=0, b=2, c=0
  id=2, a=1, b=0, c=0

Table Two (id here should read tableTwoId to match above query)
  id=1, d=3
  id=2, d=1

My first iteration used SUM(d)/COUNT(*) but division is messy so I’m currently using MIN(d). What would be a better way to write this query?

Answer 1

Try this:

SELECT
  tableOneId,
  tableOneData,
  d AS tableTwoData
FROM tableTwo
JOIN (select tableOneId, sum(a + b + c) AS tableOneData
      from tableone
      group by 1) x ON tableOneId = tableTwoId
where tableOneData <> d;

This will return all rows that have incorrect data in table 2.