Add Compute Median Instead of Average as a PostgreSQL TIL

README.md

@@ -10,7 +10,7 @@ working across different projects via [VisualMode](https://www.visualmode.dev/).
 
 For a steady stream of TILs, [sign up for my newsletter](https://visualmode.kit.com/newsletter).
 
-_1735 TILs and counting..._
+_1736 TILs and counting..._
 
 See some of the other learning resources I work on:
 
@@ -863,6 +863,7 @@ If you've learned something here, support my efforts writing daily TILs by
 - [Clear The Screen In psql](postgres/clear-the-screen-in-psql.md)
 - [Clear The Screen In psql (2)](postgres/clear-the-screen-in-psql-2.md)
 - [Compute Hashes With pgcrypto](postgres/compute-hashes-with-pgcrypto.md)
+- [Compute Median Instead Of Average](postgres/compute-median-instead-of-average.md)
 - [Compute The Levenshtein Distance Of Two Strings](postgres/compute-the-levenshtein-distance-of-two-strings.md)
 - [Compute The md5 Hash Of A String](postgres/compute-the-md5-hash-of-a-string.md)
 - [Concatenate Strings With A Separator](postgres/concatenate-strings-with-a-separator.md)

postgres/compute-median-instead-of-average.md

@@ -0,0 +1,44 @@
+# Compute Median Instead Of Average
+
+One of the first aggregate functions we might use in PostgreSQL, besides `sum`,
+is `avg`.
+
+```sql
+select avg(book_count) as average_books_read
+from (
+  select users.id, count(books.id) as book_count
+  from users
+  left join books
+    on books.user_id = users.id
+  where books.read_in_year = 2025
+  group by users.id
+) as user_book_counts;
+```
+
+This computes the average of the set of values which sums them all up
+and divides by the count. The average (maybe you've heard this also called the
+_mean_) is not always the best way to understand data, especially when there are
+outliers.
+
+Instead, we might want to compute the _median_ value of our set of data. There
+is no easily identifiable `median` aggregate function. Instead, we can use
+`percentile_cont` with a value of `0.5`. This gets us the 50th percentile of our
+set of data which is the definition of the _median_.
+
+```sql
+select percentile_cont(0.5) within group (
+  order by book_count
+) as median_books_read
+from (
+  select users.id, count(books.id) as book_count
+  from users
+  left join books on books.user_id = users.id and books.read_in_year = 2025
+  group by users.id
+) as user_book_counts;
+```
+
+The full syntax for `percentile_cont` is `percentile_cong(precision) within
+group (order by ...)` because this is an aggregiate that has to work with an
+ordered-set of data.
+
+[source](https://www.postgresql.org/docs/current/functions-aggregate.html)