DENG-2492 Create new GA4 derived table: blogs_sessions_v2 (#5018)

* DENG-2492 initial commit for new table blogs_sessions_v2

* DENG-2492 wrap keywords with backticks

sql/moz-fx-data-marketing-prod/ga/blogs_sessions/view.sql

@@ -0,0 +1,7 @@
+CREATE OR REPLACE VIEW
+  `moz-fx-data-marketing-prod.ga.blogs_sessions`
+AS
+SELECT
+  *
+FROM
+  `moz-fx-data-marketing-prod.ga_derived.blogs_sessions_v2`

sql/moz-fx-data-marketing-prod/ga_derived/blogs_sessions_v2/metadata.yaml

@@ -0,0 +1,21 @@
+friendly_name: Blogs Sessions V2
+description: |-
+  Intermediate table containing normalized sessions for blog.mozilla.org, sourced from Google Analytics 4 (GA4)
+owners:
+- kwindau@mozilla.com
+labels:
+  incremental: true
+  owner1: kwindau@mozilla.com
+scheduling:
+  dag_name: bqetl_google_analytics_derived_ga4
+bigquery:
+  time_partitioning:
+    type: day
+    field: date
+    require_partition_filter: true
+    expiration_days: null
+  clustering:
+    fields:
+    - country
+references: {}
+deprecated: false

sql/moz-fx-data-marketing-prod/ga_derived/blogs_sessions_v2/query.sql

@@ -0,0 +1,205 @@
+--Get all page views with the page location, and a flag for whether it was an entrance or not to the session
+WITH all_page_views AS (
+  SELECT
+    PARSE_DATE('%Y%m%d', event_date) AS `date`,
+    event_timestamp,
+    user_pseudo_id || '-' || CAST(
+      (
+        SELECT
+          `value`
+        FROM
+          UNNEST(event_params)
+        WHERE
+          key = 'ga_session_id'
+        LIMIT
+          1
+      ).int_value AS STRING
+    ) AS visit_identifier,
+    device.category AS device_category,
+    device.operating_system AS operating_system,
+    device.web_info.browser AS browser,
+    device.language AS `language`,
+    geo.country AS country,
+    collected_traffic_source.manual_source AS source,
+    collected_traffic_source.manual_medium AS medium,
+    collected_traffic_source.manual_campaign_name AS campaign,
+    collected_traffic_source.manual_content AS content,
+    (
+      SELECT
+        `value`
+      FROM
+        UNNEST(event_params)
+      WHERE
+        key = 'page_location'
+      LIMIT
+        1
+    ).string_value AS page_location,
+    (
+      SELECT
+        `value`
+      FROM
+        UNNEST(event_params)
+      WHERE
+        key = 'entrances'
+      LIMIT
+        1
+    ).int_value AS is_entrance
+  FROM
+    `moz-fx-data-marketing-prod.analytics_314399816.events_*`
+  WHERE
+    _TABLE_SUFFIX = FORMAT_DATE('%Y%m%d', @submission_date)
+    AND event_name = 'page_view'
+),
+--Filter to entrance pages only, and then filter to ensure only 1 entrance page per session
+--Theoretically Google should always only send 1 per session, but in case there is ever more than 1, which happens occasionally
+entrance_page_views_only AS (
+  SELECT
+    `date`,
+    visit_identifier,
+    device_category,
+    operating_system,
+    browser,
+    `language`,
+    country,
+    source,
+    medium,
+    campaign,
+    content,
+    REGEXP_REPLACE(
+      SPLIT(page_location, '?')[SAFE_OFFSET(0)],
+      '^https://blog.mozilla.org',
+      ''
+    ) AS page_path,
+    SPLIT(
+      REGEXP_REPLACE(SPLIT(page_location, '?')[SAFE_OFFSET(0)], '^https://blog.mozilla.org', ''),
+      '/'
+    )[SAFE_OFFSET(1)] AS blog,
+    SPLIT(
+      REGEXP_REPLACE(SPLIT(page_location, '?')[SAFE_OFFSET(0)], '^https://blog.mozilla.org', ''),
+      '/'
+    )[SAFE_OFFSET(2)] AS subblog
+  FROM
+    all_page_views
+  WHERE
+    is_entrance = 1
+  QUALIFY
+    ROW_NUMBER() OVER (PARTITION BY visit_identifier ORDER BY event_timestamp ASC) = 1
+),
+staging AS (
+  SELECT
+    epvo.date,
+    epvo.visit_identifier,
+    epvo.device_category,
+    epvo.operating_system,
+    epvo.browser,
+    epvo.language,
+    epvo.country,
+    epvo.source,
+    epvo.medium,
+    epvo.campaign,
+    epvo.content,
+    epvo.blog,
+    epvo.subblog,
+    COUNT(DISTINCT(visit_identifier)) AS sessions
+  FROM
+    entrance_page_views_only epvo
+  GROUP BY
+    epvo.date,
+    epvo.visit_identifier,
+    epvo.device_category,
+    epvo.operating_system,
+    epvo.browser,
+    epvo.language,
+    epvo.country,
+    epvo.source,
+    epvo.medium,
+    epvo.campaign,
+    epvo.content,
+    epvo.blog,
+    epvo.subblog
+)
+SELECT
+  `date`,
+  visit_identifier,
+  device_category,
+  operating_system,
+  browser,
+  `language`,
+  country,
+  source,
+  medium,
+  campaign,
+  content,
+  CASE
+    WHEN blog LIKE "press%"
+      THEN "press"
+    WHEN blog = 'firefox'
+      THEN 'The Firefox Frontier'
+    WHEN blog = 'netPolicy'
+      THEN 'Open Policy & Advocacy'
+    WHEN LOWER(blog) = 'internetcitizen'
+      THEN 'Internet Citizen'
+    WHEN blog = 'futurereleases'
+      THEN 'Future Releases'
+    WHEN blog = 'careers'
+      THEN 'Careers'
+    WHEN blog = 'opendesign'
+      THEN 'Open Design'
+    WHEN blog = ""
+      THEN "Blog Home Page"
+    WHEN LOWER(blog) IN (
+        'blog',
+        'addons',
+        'security',
+        'opendesign',
+        'nnethercote',
+        'thunderbird',
+        'community',
+        'l10n',
+        'theden',
+        'webrtc',
+        'berlin',
+        'webdev',
+        'services',
+        'tanvi',
+        'laguaridadefirefox',
+        'ux',
+        'fxtesteng',
+        'foundation-archive',
+        'nfroyd',
+        'sumo',
+        'javascript',
+        'page',
+        'data'
+      )
+      THEN LOWER(blog)
+    ELSE 'other'
+  END AS blog,
+  CASE
+    WHEN blog = "firefox"
+      AND subblog IN ('ru', 'pt-br', 'pl', 'it', 'id', 'fr', 'es', 'de')
+      THEN subblog
+    WHEN blog = "firefox"
+      THEN "Main"
+    WHEN blog LIKE "press-%"
+      AND blog IN (
+        'press-de',
+        'press-fr',
+        'press-es',
+        'press-uk',
+        'press-pl',
+        'press-it',
+        'press-br',
+        'press-nl'
+      )
+      THEN blog
+    WHEN blog LIKE "press%"
+      THEN "Main"
+    WHEN blog = 'internetcitizen'
+      AND subblog IN ('de', 'fr')
+      THEN subblog
+    ELSE "Main"
+  END AS subblog,
+  `sessions`
+FROM
+  staging

sql/moz-fx-data-marketing-prod/ga_derived/blogs_sessions_v2/schema.yaml

@@ -0,0 +1,57 @@
+fields:
+- mode: NULLABLE
+  name: date
+  type: DATE
+  description: Date of the visit
+- mode: NULLABLE
+  name: visit_identifier
+  type: STRING
+  description: Visit Identifier - Uniquely identifies a visit; concatenation of user_pseudo_id and ga_session_id
+- mode: NULLABLE
+  name: device_category
+  type: STRING
+  description: Device Category - The device category the visitor used to visit the site
+- mode: NULLABLE
+  name: operating_system
+  type: STRING
+  description: Operating System - The operating system the visitor used to visit the site
+- mode: NULLABLE
+  name: browser
+  type: STRING
+  description: Browser - The browser the visiting device was using when it visited the site
+- mode: NULLABLE
+  name: language
+  type: STRING
+  description: Language - The language the visiting device was using when it visited the site
+- mode: NULLABLE
+  name: country
+  type: STRING
+  description: Country - The country from which events were reported, based on IP address
+- mode: NULLABLE
+  name: source
+  type: STRING
+  description: Source - Referring partner domain
+- mode: NULLABLE
+  name: medium
+  type: STRING
+  description: Medium - Category of the source, such as 'organic' for a search engine
+- mode: NULLABLE
+  name: campaign
+  type: STRING
+  description: Campaign - Identifier for the marketing campaign
+- mode: NULLABLE
+  name: content
+  type: STRING
+  description: Content - Indicates the particular link within a campaign
+- mode: NULLABLE
+  name: blog
+  type: STRING
+  description: Blog
+- mode: NULLABLE
+  name: subblog
+  type: STRING
+  description: Sub-Blog
+- mode: NULLABLE
+  name: sessions
+  type: INT64
+  description: Number of Sessions