backfill/2024_10_10_kwindau_clients_first_seen_v2/metadata.yaml

friendly_name: Clients First Seen
description: |-
  Picks out just the first observations per client based on
  the main, new_profile and first_shutdown pings.
  - All attributes are retrieved from the ping that reports
  the first_seen_date, using the earliest timestamp and respecting nulls.
  - In the case of different ping types received in the same timestamp,
  new_profile ping is prioritized.
  - Only the second_seen_date and client's reported pings are updated
  daily when NULL.
  This table should be accessed through the user-facing view
  `telemetry.clients_first_seen`.
  See caveats for analysis and how to use this table:
  https://docs.google.com/document/d/1qMkTUIruTOxTGRj7S60nEsyNjGHYAuVJ51-53NZxnd4/edit?usp=sharing

  For caveats and how to use this data please refer to
  https://docs.google.com/document/d/1qMkTUIruTOxTGRj7S60nEsyNjGHYAuVJ51-53NZxnd4/edit#heading=h.3110lg9vb11d

  The initialization command creates the table and then inserts data
  in parallel for each sample_id. Usage:
  `bqetl query initialize sql/moz-fx-data-shared-prod/telemetry_derived/clients_first_seen_v2/query.sql`
owners:
- lvargas@mozilla.com
labels:
  application: firefox
  incremental: true
  schedule: daily
  depends_on_past: true
  owner1: lvargas
scheduling:
  dag_name: bqetl_analytics_tables
  task_name: clients_first_seen_v2
  depends_on_past: true
  date_partition_parameter: null
  parameters:
  - submission_date:DATE:{{ds}}
bigquery:
  time_partitioning:
    type: day
    field: first_seen_date
    require_partition_filter: false
    expiration_days: null
  clustering:
    fields:
    - normalized_channel
    - sample_id
    - country
    - normalized_os
references: {}