From dc60b040f517b053d6513299dc179e10e674441b Mon Sep 17 00:00:00 2001
From: Leli <33942105+lelilia@users.noreply.github.com>
Date: Thu, 29 Feb 2024 19:24:57 +0100
Subject: [PATCH] DENG-2918-UDFs for URL parsing (#5141)

* DENG-2918-UDFs for URL parsing

* removing get_host because net.host() exists
---
 sql/mozfun/utils/get_url_path/README.md     |  4 ++++
 sql/mozfun/utils/get_url_path/metadata.yaml |  2 ++
 sql/mozfun/utils/get_url_path/udf.sql       | 21 +++++++++++++++++++++
 3 files changed, 27 insertions(+)
 create mode 100644 sql/mozfun/utils/get_url_path/README.md
 create mode 100644 sql/mozfun/utils/get_url_path/metadata.yaml
 create mode 100644 sql/mozfun/utils/get_url_path/udf.sql

diff --git a/sql/mozfun/utils/get_url_path/README.md b/sql/mozfun/utils/get_url_path/README.md
new file mode 100644
index 0000000000..e4d300a80a
--- /dev/null
+++ b/sql/mozfun/utils/get_url_path/README.md
@@ -0,0 +1,4 @@
+This UDF extracts path from a URL string.
+
+The path is everything after the host and before parameters.
+This function returns "/" if there is no path.
\ No newline at end of file
diff --git a/sql/mozfun/utils/get_url_path/metadata.yaml b/sql/mozfun/utils/get_url_path/metadata.yaml
new file mode 100644
index 0000000000..ad56bad597
--- /dev/null
+++ b/sql/mozfun/utils/get_url_path/metadata.yaml
@@ -0,0 +1,2 @@
+description: Extract the Path from a URL
+friendly_name: Get URL Path
diff --git a/sql/mozfun/utils/get_url_path/udf.sql b/sql/mozfun/utils/get_url_path/udf.sql
new file mode 100644
index 0000000000..86a347155b
--- /dev/null
+++ b/sql/mozfun/utils/get_url_path/udf.sql
@@ -0,0 +1,21 @@
+CREATE OR REPLACE FUNCTION utils.get_url_path(url STRING)
+RETURNS STRING AS (
+  "/" || COALESCE(
+    REGEXP_EXTRACT(REPLACE(REPLACE(url, "https://", ""), "http://", ""), r"\/([^&?#]*)"),
+    ""
+  )
+);
+
+-- Tests
+SELECT
+  mozfun.assert.equals("/path", utils.get_url_path("https://some-url.com/path")),
+  mozfun.assert.equals("/path", utils.get_url_path("http://some-url.com/path?more")),
+  mozfun.assert.equals("/path", utils.get_url_path("http://some-url.com/path#more")),
+  mozfun.assert.equals("/path", utils.get_url_path("http://some-url.com/path?more&utm=123")),
+  mozfun.assert.equals(
+    "/path/with/multiple/slashes",
+    utils.get_url_path("http://some-url.com/path/with/multiple/slashes?more")
+  ),
+  mozfun.assert.equals("/", utils.get_url_path("https://some-url.com")),
+  mozfun.assert.equals("/", utils.get_url_path("https://some-url.com/")),
+  mozfun.assert.equals("/path", utils.get_url_path("some-url.com/path"))