Add UDFs to generate regular expressions to correctly match event properties in `events_daily` tables (#4272)

* Add `event_analysis.event_property_value_to_match_string` UDF.

* Add `event_analysis.event_property_index_to_match_string` UDF.

sql/mozfun/event_analysis/event_property_index_to_match_string/metadata.yaml

@@ -0,0 +1,6 @@
+---
+description: >
+  Given an event index and property index from an `event_types` table, returns a regular expression
+  to match corresponding events within an `events_daily` table's `events` string that aren't missing
+  the specified property.
+friendly_name: Event Analysis Event Property Index To Match String

sql/mozfun/event_analysis/event_property_index_to_match_string/udf.sql

@@ -0,0 +1,34 @@
+CREATE OR REPLACE FUNCTION event_analysis.event_property_index_to_match_string(
+  event_index STRING,
+  property_index INTEGER
+)
+RETURNS STRING AS (
+  CONCAT(
+    -- Double-quote characters are used in `events_daily.events` strings to indicate missing properties.
+    '[^",]',
+    -- Event property values are stored in `events_daily.events` strings in reverse order,
+    -- so we expect the Nth property value to be followed by N-1 other property values.
+    IF(property_index > 1, CONCAT('[^,]{', (property_index - 1), '}'), ''),
+    event_analysis.event_index_to_match_string(event_index)
+  )
+);
+
+SELECT
+  assert.equals(r'[^",]\Qe\E,', event_analysis.event_property_index_to_match_string('e', 1)),
+  assert.equals(r'[^",][^,]{1}\Qe\E,', event_analysis.event_property_index_to_match_string('e', 2)),
+  assert.equals(
+    'pe,',
+    REGEXP_EXTRACT('""pe,', event_analysis.event_property_index_to_match_string('e', 1))
+  ),
+  assert.equals(
+    CAST(NULL AS STRING),
+    REGEXP_EXTRACT('""pe,', event_analysis.event_property_index_to_match_string('e', 2))
+  ),
+  assert.equals(
+    CAST(NULL AS STRING),
+    REGEXP_EXTRACT('"p"e,', event_analysis.event_property_index_to_match_string('e', 1))
+  ),
+  assert.equals(
+    'p"e,',
+    REGEXP_EXTRACT('"p"e,', event_analysis.event_property_index_to_match_string('e', 2))
+  ),

sql/mozfun/event_analysis/event_property_value_to_match_string/metadata.yaml

@@ -0,0 +1,5 @@
+---
+description: >
+  Given an event index, property index, and property value from an `event_types` table, returns a
+  regular expression to match corresponding events within an `events_daily` table's `events` string.
+friendly_name: Event Analysis Event Property Value To Match String

sql/mozfun/event_analysis/event_property_value_to_match_string/udf.sql

@@ -0,0 +1,41 @@
+CREATE OR REPLACE FUNCTION event_analysis.event_property_value_to_match_string(
+  event_index STRING,
+  property_index INTEGER,
+  property_value STRING
+)
+RETURNS STRING AS (
+  CONCAT(
+    event_analysis.escape_metachars(property_value),
+    -- Event property values are stored in `events_daily.events` strings in reverse order,
+    -- so we expect the Nth property value to be followed by N-1 other property values.
+    IF(property_index > 1, CONCAT('[^,]{', (property_index - 1), '}'), ''),
+    event_analysis.event_index_to_match_string(event_index)
+  )
+);
+
+SELECT
+  assert.equals(r'\Qp\E\Qe\E,', event_analysis.event_property_value_to_match_string('e', 1, 'p')),
+  assert.equals(
+    r'\Qp\E[^,]{1}\Qe\E,',
+    event_analysis.event_property_value_to_match_string('e', 2, 'p')
+  ),
+  assert.equals(
+    CAST(NULL AS STRING),
+    event_analysis.event_property_value_to_match_string('e', 1, NULL)
+  ),
+  assert.equals(
+    'pe,',
+    REGEXP_EXTRACT('""pe,', event_analysis.event_property_value_to_match_string('e', 1, 'p'))
+  ),
+  assert.equals(
+    CAST(NULL AS STRING),
+    REGEXP_EXTRACT('""pe,', event_analysis.event_property_value_to_match_string('e', 2, 'p'))
+  ),
+  assert.equals(
+    CAST(NULL AS STRING),
+    REGEXP_EXTRACT('"p"e,', event_analysis.event_property_value_to_match_string('e', 1, 'p'))
+  ),
+  assert.equals(
+    'p"e,',
+    REGEXP_EXTRACT('"p"e,', event_analysis.event_property_value_to_match_string('e', 2, 'p'))
+  ),