feat: Use LowCardinality for smaller columns

Generally for columns with less than 10k unique values this provides a noticable performance benefit.

models/base/xapi_events_all_parsed.sql

@@ -13,7 +13,7 @@
 
 select
     event_id as event_id,
-    JSON_VALUE(event::String, '$.verb.id') as verb_id,
+    toLowCardinality(JSON_VALUE(event::String, '$.verb.id')) as verb_id,
     COALESCE(
         NULLIF(JSON_VALUE(event::String, '$.actor.account.name'), ''),
         NULLIF(JSON_VALUE(event::String, '$.actor.mbox'), ''),
@@ -23,7 +23,7 @@ select
     -- If the contextActivities parent is a course, use that. It can be a "course"
     -- type, or a "cmi.interaction" type for multiple question problem submissions.
     -- Otherwise use the object id for the course id.
-    multiIf(
+    toLowCardinality(multiIf(
         -- If the contextActivities parent is a course, use that
         JSON_VALUE(
             event::String, '$.context.contextActivities.parent[0].definition.type'
@@ -37,10 +37,11 @@ select
         JSON_VALUE(event::String, '$.context.contextActivities.grouping[0].id'),
         -- Otherwise use the object id
         JSON_VALUE(event::String, '$.object.id')
-    ) as course_id,
-    coalesce(
+    )) as course_id,
+    toLowCardinality(splitByString('/', course_id)[-1]) as course_key,
+    toLowCardinality(coalesce(
         get_org_from_course_url(course_id), get_org_from_ccx_course_url(course_id), ''
-    ) as org,
+    )) as org,
     emission_time as emission_time,
     event::String as event
 from {{ source("xapi", "xapi_events_all") }}

models/completion/completion_events.sql

@@ -15,7 +15,7 @@ select
     CAST(emission_time, 'DateTime') as emission_time,
     actor_id,
     object_id,
-    splitByString('/', course_id)[-1] as course_key,
+    course_key,
     org,
     verb_id,
     JSON_VALUE(

models/enrollment/enrollment_events.sql

@@ -15,13 +15,13 @@ select
     cast(emission_time as DateTime) as emission_time,
     actor_id,
     object_id,
-    splitByString('/', course_id)[-1] as course_key,
+    course_key,
     org,
     verb_id,
-    JSON_VALUE(
+    toLowCardinality(JSON_VALUE(
         event,
         '$.object.definition.extensions."https://w3id.org/xapi/acrossx/extensions/type"'
-    ) as enrollment_mode
+    )) as enrollment_mode
 from {{ ref("xapi_events_all_parsed") }}
 where
     verb_id in (

models/forum/forum_events.sql

@@ -14,7 +14,7 @@ select
     event_id,
     CAST(emission_time, 'DateTime') as emission_time,
     org,
-    splitByString('/', course_id)[-1] as course_key,
+    course_key,
     object_id,
     actor_id,
     verb_id

models/grading/grading_events.sql

@@ -16,7 +16,7 @@ select
     CAST(emission_time, 'DateTime') as emission_time,
     actor_id,
     object_id,
-    splitByString('/', course_id)[-1] as course_key,
+    course_key,
     org,
     verb_id,
     JSONExtractFloat(event, 'result', 'score', 'scaled') as scaled_score

models/navigation/navigation_events.sql

@@ -15,7 +15,7 @@ select
     cast(emission_time as DateTime) as emission_time,
     actor_id,
     splitByString('/xblock/', object_id)[-1] as block_id,
-    splitByString('/', course_id)[-1] as course_key,
+    course_key,
     org,
     verb_id,
     JSONExtractString(event, 'object', 'definition', 'type') as object_type,

models/problems/problem_events.sql

@@ -15,7 +15,7 @@ select
     cast(emission_time as DateTime) as emission_time,
     actor_id,
     object_id,
-    splitByString('/', course_id)[-1] as course_key,
+    course_key,
     org,
     verb_id,
     JSON_VALUE(event, '$.result.response') as responses,
@@ -25,7 +25,7 @@ select
         cast(JSON_VALUE(event, '$.result.success') as Bool),
         false
     ) as success,
-    JSON_VALUE(event, '$.object.definition.interactionType') as interaction_type,
+    toLowCardinality(JSON_VALUE(event, '$.object.definition.interactionType')) as interaction_type,
     if(
         verb_id = 'https://w3id.org/xapi/acrossx/verbs/evaluated',
         cast(

models/video/video_playback_events.sql

@@ -15,7 +15,7 @@ select
     CAST(emission_time, 'DateTime') as emission_time,
     actor_id,
     object_id,
-    splitByString('/', course_id)[-1] as course_key,
+    course_key,
     org,
     verb_id,
     ceil(

models/video/video_transcript_events.sql

@@ -14,7 +14,7 @@ select
     event_id,
     CAST(emission_time, 'DateTime') as emission_time,
     org,
-    splitByString('/', course_id)[-1] as course_key,
+    course_key,
     splitByString('/xblock/', object_id)[2] as video_id,
     actor_id,
     JSONExtractBool(
@@ -25,7 +25,7 @@ select
     ) as cc_enabled
 from {{ ref("xapi_events_all_parsed") }}
 where
-    verb_id in ('http://adlnet.gov/expapi/verbs/interacted')
+    verb_id = 'http://adlnet.gov/expapi/verbs/interacted'
     and JSONHas(
         event,
         'result',