Question

I have two DB tabels which form a parent child relationship from Planung to Aufgabe:

Planung:

CREATE TABLE `planung` (
  `id` bigint(20) NOT NULL AUTO_INCREMENT,
  `Bezeichnung` varchar(255) DEFAULT NULL,
   -- lots of ohter columns
  PRIMARY KEY (`id`),
) ENGINE=InnoDB

Aufgabe:

CREATE TABLE `aufgabe` (
  `id` bigint(20) NOT NULL AUTO_INCREMENT,
  `planung_id` bigint(20) DEFAULT NULL, -- foreign key to Planung.id
  `Nummer` int(11) DEFAULT NULL,
  PRIMARY KEY (`id`),
  -- lots of ohter columns
) ENGINE=InnoDB 

I’m looking for a query which gives me all Planung.id and Aufgabe.ID for all duplicate Nummer per Planung. Or in other words: For every Planung Aufgabe.Nummer must be unique, I want to check whehter this is really the case in my DB (I know it isn’t).

Answer 1

SELECT planung_id, GROUP_CONCAT(id) AS aufgabe_id, Nummer, COUNT(1) as num_duplicates
FROM aufgabe
GROUP BY planung_id, Nummer
HAVING COUNT(1) > 1

This query gives all planung_id which has duplicate Nummer and displays them like:

planung_id aufgabe_id Nummer num_duplicates
1          2,5,8      1      3

Which means, for Planung 1, there exists three Aufgabe with Nummer 1, and they are 2,5 and 8.

Edit: Shamelessly stolen from @dgw‘s comment:
Once you have run this query, and corrected all the duplicates. Add a unique index to aufgabe {planung_id, Nummer} to ensure that the database maintains this constraint:

ALTER TABLE aufgabe 
  ADD CONSTRAINT UNIQUE uq_planung_id_and_nummer (planung_id, Nummer)