transform pull request - switch to single column PK

Vertica does not support multi-column PK in incremental loads

I created an issue for that:
vertica/dbt-vertica#109

data_pipeline/models/github/pull_requests/pull_requests.sql

@@ -7,7 +7,7 @@
     {'columns': ['created_at'], 'unique': false}
   ],
   materialized='incremental',
-  unique_key=['pull_request_number', 'repo_id']
+  unique_key='id'
 ) }}
 
 with using_clause as (
@@ -22,7 +22,7 @@ updates as (
   select *
   from using_clause
   {% if is_incremental() %}
-    where (pull_request_number, repo_id) in ( select pull_request_number, repo_id from {{ this }} )
+    where id in ( select id from {{ this }} )
   {% else %}
     -- No updates when doing full load
     where 1 = 0
@@ -33,7 +33,7 @@ inserts as (
   select *
   from using_clause
   {% if is_incremental() %}
-    where (pull_request_number, repo_id) not in ( select pull_request_number, repo_id from {{ this }} )
+    where id not in ( select id from {{ this }} )
   {% endif %}
 ),
 
@@ -47,6 +47,10 @@ repos as (
 
 final as (
     select
+      -- Have to SELECT "id" here for incremental processing
+      -- It is not defined in schema.yml, it is not exposed to GoodData
+      id,
+      -- replace useless internal PR ID with something meaningful
       repos.repo_name || '/' || p.pull_request_number as pull_request_id,
       p.pull_request_number,
       p.pull_request_url,

data_pipeline/models/github/pull_requests/pull_requests_extract_json.sql

@@ -6,14 +6,15 @@
     {'columns': ['created_at'], 'unique': false}
   ],
   materialized='incremental',
-  unique_key=['pull_request_number', 'repo_id']
+  unique_key='id'
 ) }}
 
 -- Helper step, materialize extracted JSON fields first and then JOIN it with other tables
 -- Incremental mode
 
 with using_clause as (
   select
+    id,
     number as pull_request_number,
     html_url as pull_request_url,
     title as pull_request_title,
@@ -24,7 +25,6 @@ with using_clause as (
     merged_at,
     closed_at,
     {{ extract_json_value('user', 'id', 'user_id', 'INT') }}
-    --CAST(json_extract_path_text(to_json("{{ get_db_entity_name('user') }}"), 'id') as INT) as user_id
   from {{ var("input_schema_github") }}.pull_requests
   {% if is_incremental() %}
     where created_at > ( select max(created_at) from {{ this }} )
@@ -35,7 +35,7 @@ updates as (
   select *
   from using_clause
   {% if is_incremental() %}
-    where (pull_request_number, repo_id) in ( select pull_request_number, repo_id from {{ this }} )
+    where id in ( select id from {{ this }} )
   {% else %}
     -- No updates when doing full load
     where 1 = 0
@@ -46,7 +46,7 @@ inserts as (
   select *
   from using_clause
   {% if is_incremental() %}
-    where (pull_request_number, repo_id) not in ( select pull_request_number, repo_id from {{ this }} )
+    where id not in ( select id from {{ this }} )
   {% endif %}
 ),