xnotes.md

Implementaiton Notes

Model shared between dialects

• imdb.model.lookml, imdb_spark.model.lookml, imdb.base.lookml
• Runs both Spark and Redshift
• Uses a base file with a base explore, each model extends
• Spark has really poorly optimized Left Joins, so we inner join
• Redshift doesn’t support lists
• liquid to conditionally compile based on dialect - movie_budget.view.lookml

Reverse Join Pattern

• Single Base View (title) ~ 3M
• Titles have Cast Memebers, Production Companies, Genres, Languages and Keywords
• Each of these attributes are one_to_many joins
• Many of the attributes we join twice so we can see co-occurances

Polymorphic Table

• Title is both tv episode and tv series.

Break up movie_info with derived tables.