Question

I’m looking for a query that allows me to “flatten” rows with the same ID:

• Values from the second column are used as the new tables column names
• Values from the third column determine the values in the new table.

As an example:

A table looks like this.

ID       year     event
------------------------
Guid1    2005     A
Guid1    2010     B
Guid2    2020     C

I want it to look like..

ID       2005      2010     2020
Guid1    A         B        
Guid2                       C

Any ideas on how to do this?

Answer 1

I assume (id, year) combination is unique? Otherwise not sure what output you expect. If unique, it looks like this:

SELECT
  id,
  MIN(IF(year = 2005, event, NULL)) AS '2005',
  MIN(IF(year = 2010, event, NULL)) AS '2010',
  MIN(IF(year = 2020, event, NULL)) AS '2020'
FROM
  t
GROUP BY
  id

The above means you’ll have to repeat the MIN(IF(...)) statement once for each desired year. There is no way in SQL to do this automatically: SQL does not mix data and metadata. You cannot dynamically specify columns in a query. Either you use “*” or you specify the exact columns.