Question

The table MyTable contains 4 columns.

• id, phone columns are unique.
• id, phone, zipcode are never empty.

---

id - phone -        salary   - zipcode

1  - 61730459987 - $56052.45 - 02456

2 - 21249620709 - NULL -      10023

3 - 21200830859 - $39089.28 - 10023

...

10000 - 64600830857 - $46063.72 - 03795**

I need to remove NULLs by replacing them with salary_estimates, based on information from the same zip-codes.

As for the example above, NULL at row #2 could be replaced with $39089.28 (we can find from the table that another person with the same zip-code 10023 earns $39089.28, so it’s OK to update NULL for #2 with $39089.28).
So, after we run this query, the number of empty cells could be reduced (although not up to zero).

Could anybody suggest an update query for this problem?

Answer 1

UPDATE
      MyTable AS tu
  JOIN 
      ( SELECT zipcode
             , AVG(salary) AS SalaryEstimate          --- AVG, MAX, MIN
                                                      --- your choice
        FROM MyTable 
        WHERE salary IS NOT NULL
        GROUP BY zipcode
      ) AS te
    ON te.zipcode = tu.zipcode
SET 
      tu.salary = te.SalaryEstimate
WHERE 
      tu.salary IS NULL