Exclude non-expert-validated plays.

[jamiefolsom]

This filters out play which are not expert_validated when the warehouse script is run.

It occurs to me that there could be a related problem in which existing tickets sales, or registers, are still referring to these non-expert validated plays, and which therefore need to be re-mapped to their expert validated equivalents.

I will (and you can) test the PR without affecting the staging or production databases, run the warehouse rake task on this heroku app: https://dashboard.heroku.com/apps/cfrp-staging-pr-3

[jamiefolsom]

Update -- testing this PR on the Review app, indeed the rake task aborts because of the same foreign key constraint issue.

ActiveRecord::InvalidForeignKey: PG::ForeignKeyViolation: ERROR:  insert or update on table "sales_facts" violates foreign key constraint "sales_facts_play_2_id_fkey"
DETAIL:  Key (play_2_id)=(856) is not present in table "play_dim".
: UPDATE warehouse.sales_facts SET play_2_id = (SELECT min(play_id) FROM register_plays WHERE ordering = 2 AND register_plays.register_id = sales_facts.register_id)

So a next step would be to find out which registers are still referencing non-validated plays, and remap them to validated ones.

[forestofarden]

Hi Jamie - It may help to know that the warehouse plays_dim table uses the same ids as the original plays table.

Here are the bad plays:

cfrp_development=# select distinct(play_dim.id) as play_id from warehouse.sales_facts, warehouse.play_dim where play_dim.id in (play_1_id, play_2_id, play_3_id, play_4_id) and acts is null;
 play_id 
---------
     856
    5677
    5878
    5917
(4 rows)

And, the bad register entries:

cfrp_development=# select registers.id, date, ordering, plays.id, title from register_plays join registers on (registers.id = register_id) join plays on(plays.id = play_id) where play_id in (856,5917,5878,5677) and verification_state_id in (1,6) order by date, ordering;
  id   |    date    | ordering |  id  |          title           
-------+------------+----------+------+--------------------------
 24629 | 1684-11-23 |        2 | 5917 | Danse de Gigogne
 24654 | 1684-12-19 |        2 | 5917 | Danse de Gigogne
 34950 | 1688-10-27 |        2 | 5677 | Le Désespoir extravagant
 17370 | 1746-04-24 |        3 |  856 | 
 17377 | 1746-05-01 |        3 |  856 | 
 21142 | 1759-05-23 |        1 | 5878 | Suivante généreuse (La)
 21144 | 1759-05-26 |        1 | 5878 | Suivante généreuse (La)
  3289 | 1770-02-19 |        2 |  856 | 
  3295 | 1770-02-25 |        2 |  856 | 
  3407 | 1770-07-25 |        2 |  856 | 
  3526 | 1770-12-02 |        2 |  856 | 
  3551 | 1770-12-30 |        2 |  856 | 
  4172 | 1773-01-23 |        2 |  856 | 
  2954 | 1774-02-10 |        2 |  856 | 
  2957 | 1774-02-13 |        2 |  856 | 
  2959 | 1774-02-15 |        2 |  856 | 
  2964 | 1774-02-20 |        2 |  856 | 
  2967 | 1774-02-23 |        2 |  856 | 
  2971 | 1774-02-27 |        2 |  856 | 
  2983 | 1774-03-13 |        2 |  856 | 
(20 rows)

Hope this helps! C

[jamiefolsom]

@christopheryork -- Yep, that makes sense. I will request an update from Sara as to which plays those registers should point to instead. Thanks for confirming, and for those queries!

[forestofarden]

Sounds good. Good luck!