SQL query, need help understanding - sql

Question

I' am new to SQL and seem to struggle when I come across harder queries. I cannot seem to understand how to write these SQL queries from start to finish. I get lost on where to start and how to build these up one at time and combine together. Below is a questions I was asked followed by the solution. Would anyone help me walk through a way of looking at this that might help me understand to get the same results?
Write a query to return the shortest movie from each category.
The order of your results doesn't matter.
If there are ties, return just one of them.
Return the following columns: film_id, title, length, category, row_num
WITH movie_ranking AS (
SELECT
F.film_id,
F.title,
F.length,
C.name category,
ROW_NUMBER() OVER(PARTITION BY C.name ORDER BY F.length) row_num
FROM film F
INNER JOIN film_category FC
ON FC.film_id = F.film_id
INNER JOIN category C
ON C.category_id = FC.category_id
)
SELECT
film_id,
title,
length,
category,
row_num
FROM movie_ranking
WHERE row_num = 1
;
SELECT
film_id,
title,
length,
category,
row_num
FROM (
SELECT
F.film_id,
F.title,
F.length,
C.name category,
ROW_NUMBER() OVER(PARTITION BY C.name ORDER BY F.length) row_num
FROM film F
INNER JOIN film_category FC
ON FC.film_id = F.film_id
INNER JOIN category C
ON C.category_id = FC.category_id
) X
WHERE row_num = 1
;

Answer 1

Lets see if this helps. First, can you follow the very SIMPLE query of "get a list of all movies and sort them by category. Within each category, sort with the shortest movies to longest.
select
c.name category,
f.length,
f.title
from
film f
JOIN film_category fc
on f.film_id = fc.film_id
join category c
on fc.category_id = c.category_id
order by
c.name,
f.length
This should be pretty simple. If not, I will go deeper. So how does this help you with partitioning and ordering, and what is this ROW_NUMBER() thing.
First, partitioning. Think of this as taking the result records that come through and pre-breaking them down into different buckets (your film categories). Simple enough, as the above is just ordering the output based on the c.name (category name descriptive text).
Now, the ORDER BY clause in the partitioning. That corresponds to your secondary sort on the film length. So the first part took all the films and put them into different "buckets" Sci-Fi, RomCom, Horror, Fiction, Action/Adventure, etc. Now the ORDER BY is saying, within each bucket, sort the data (hence the order by WITHIN each c.name). At this time, you now have 5+ buckets (each category), each with different number of movies. Some could have 4 or 5 example, others could have 30, others 18, whatever. But at least now, each category bucket has been sorted based on the length.
NOW comes the ROW_NUMBER() (partion context) row_num. ROW_NUMBER() is a Windows Function call that resets back to 0 each time a new bucket (film category) is encountered and so it takes the records now sorted by length and assigns it a sequential # 1, 2, 3, etc for however many are within each category. The final column name from the ROW_NUMBER() is "row_num" but could be assigned any column name.
So we now have 4, 5 or however many buckets, all sorted by the length of the movie IN EACH RESPECTIVE BUCKET. And each bucket as a # 1, 2, 3, etc based on available movies.
Now the query is using "CTE" (common table expression). Basically writing a query with an alias name that will be used in the immediate following actual query. So this "WITH MOVIE_RANKING" query is similar to my query but without a generic outer ORDER BY clause, but instead the partition and order within each partition.
WITH movie_ranking AS (
SELECT
F.film_id,
F.title,
F.length,
C.name category,
ROW_NUMBER() OVER(PARTITION BY C.name ORDER BY F.length) row_num
FROM film F
INNER JOIN film_category FC
ON FC.film_id = F.film_id
INNER JOIN category C
ON C.category_id = FC.category_id
)
SELECT
film_id,
title,
length,
category,
row_num
FROM movie_ranking
WHERE row_num = 1
;
The final select gets the result of the WITH Movie_Ranking alias context, so it can explicitly reference the final column name called "row_num". Since the request is to get the move with the shortest time in each category, all you care about is any record with the "row_num" = 1. Remember, each bucket was sorted, then assigned sequentially, so #1 in each bucket was the shortest length movie.
The second version of the query is just doing an in-line query from a query. Basically an in-line copy of the WITH CTE, just using the now recognized alias to keep readability simplified.