increment_key

Usage

view: my_view {
  derived_table: {
    increment_key: "created_date"
    ...
  }
}
Hierarchy
increment_key

- or -

increment_key
Default Value
None

Accepts
The name of a time-based LookML dimension

Special Rules
increment_key is supported only with persistent tables, and only for specific dialects

Definition

You can create incremental PDTs in your project if your dialect supports them. An incremental PDT is a persistent derived table (PDT) that Looker builds by appending fresh data to the table, instead of rebuilding the table in its entirety. See the Incremental PDTs documentation page for more information.

increment_key is the parameter that makes a PDT into an incremental PDT by specifying the time increment for which fresh data should be queried and appended to the PDT. In addition to the increment_key, you can optionally provide an increment_offset to specify the number of previous time periods (at the increment key's granularity) that are rebuilt to account for late-arriving data.

The increment_key for a PDT is independent of the PDT's persistence trigger. See the Incremental PDTs documentation page for some example scenarios that show the interaction of increment_key, increment_offset, and persistence strategy.

The increment_key parameter works only with supported dialects, and only with tables that have a persistence strategy, such as PDTs and aggregate tables (which are a type of PDT).

The increment_key must specify a time-based LookML dimension:

In addition, the increment_key must be:

  • A truncated absolute time, such as day, month, year, fiscal quarter, and so on. Timeframes such as day of the week are not supported.
  • A timestamp that increases predictably with new data, such as order created date. In other words, a timestamp should be used as an increment key only if the newest data added to the table also has the latest timestamp. A timestamp such as user birthday would not work as an increment key, since a birthday timestamp does not increase reliably with new users added to the table.

Creating an incremental LookML-based PDT

To make a LookML-based (native) PDT into an incremental PDT, use the increment_key parameter to specify the name of a time-based LookML dimension. The dimension must be defined in the view on which the PDT's explore_source is based.

For example, here is a view file for a PDT based on LookML, using the explore_source LookML parameter. The PDT is created from the flights Explore, which in this case is based on the flights view:

view: flights_lookml_incremental_pdt {
  derived_table: {
    indexes: ["id"]
    increment_key: "departure_date"
    increment_offset: 3
    datagroup_trigger: flights_default_datagroup
    distribution_style: all
    explore_source: flights {
      column: id {}
      column: carrier {}
      column: departure_date {}
    }
  }

  dimension: id {
    type: number
  }
  dimension: carrier {
    type: string
  }
  dimension: departure_date {
    type: date
  }
}

This table will build in its entirety the first time a query is run on it. After that, the PDT will be rebuilt in increments of one day (increment_key: departure_date), going back three days (increment_offset: 3).

The departure_date dimension is actually the date timeframe from the departure dimension group. (See the dimension_group parameter documentation page for an overview of how dimension groups work.) The dimension group and timeframe are both defined in the flights view, which is the explore_source for this PDT. Here is how the departure dimension group is defined in the flights view file:

...
  dimension_group: departure {
    type: time
    timeframes: [
      raw,
      date,
      week,
      month,
      year
    ]
    sql: ${TABLE}.dep_time ;;
  }
...

Creating an incremental SQL-based PDT

Looker suggests that you use LookML-based (native) derived tables as the base for incremental PDTs, as opposed to using SQL-based derived tables. Native derived tables inherently handle the complex logic required for incremental PDTs. SQL-based PDTs rely on manually created logic, which is prone to error when used with highly complex functionality.

To define an incremental SQL-based PDT, use increment_key and (optionally) increment_offset as you would with a LookML-based PDT. However, because SQL-based PDTs aren't based on LookML view files, there are additional requirements for making a SQL-based PDT into an incremental PDT:

  • You must base the increment key on a time-based LookML dimension that you define in the PDT's view file.
  • You must provide a {% incrementcondition %} Liquid filter in the PDT to connect the increment key to the database time column that the increment key is based on. The {% incrementcondition %} filter must specify the name of the column in your database, not a SQL alias nor the name of a dimension that is based on the column (see the following example).

The basic format for the Liquid filter is:

   WHERE {% incrementcondition %} database_table_name.database_time_column {% endincrementcondition %}

For example, here is the view file for a SQL-based PDT that is rebuilt in increments of one day (increment_key: "dep_date"), where data from the last three days will be added to the table when it is rebuilt (increment_offset: 3):

view: sql_based_incremental_date_pdt {
  derived_table: {
    datagroup_trigger: flights_default_datagroup
    increment_key: "dep_date"
    increment_offset: 3
    distribution_style: all
    sql: SELECT
        flights.id2  AS "id",
        flights.origin  AS "origin",
        DATE(flights.leaving_time )  AS "departure"
      FROM public.flights  AS flights
      WHERE {% incrementcondition %} flights.leaving_time {%  endincrementcondition %}
          ;;
  }

  dimension_group: dep {
    type: time
    timeframes: [date, week, month, year]
    datatype: date
    sql:  ${TABLE}.departure
    ;;
  }
  dimension: id {
      type: number
    }
    dimension: origin {
      type: string
  }
}

Note the following about this example:

  • The derived table is based on a SQL statement. The SQL statement creates a column in the derived table that is based on the flights.leaving_time column in the database. The column is given the alias departure.
  • The PDT's view file defines a dimension group called dep.
    • The dimension group's sql parameter indicates that the dimension group is based on the departure column in the derived table.
    • The dimension group's timeframes parameter includes date as a timeframe.
  • The derived table's increment_key uses the dep_date dimension, which is a dimension based on the date timeframe of the dep dimension group. (See the dimension_group parameter documentation page for an overview of how dimension groups work.)
  • The {% incrementcondition %} Liquid filter is used to connect the increment key to the flights.leaving_time column in the database.
    • The {% incrementcondition %} must specify the name of a TIMESTAMP column in your database (or it must evaluate to a TIMESTAMP column in your database).
    • The {% incrementcondition %} must evaluate against what is available in the FROM clause that defines your PDT, such as columns from the table that is specified in the FROM clause. The {% incrementcondition %} cannot refer to the result of the SELECT statement, such as an alias that has been given to a column in the SQL statement, or the name of a dimension that is based on the column. In this example, the {% incrementcondition %} is flights.leaving_time. Since the FROM clause specifies the flights table, the {% incrementcondition %} can refer to columns from the flights table.
    • The {% incrementcondition %} must point to the same database column that is used for the increment key. In this example, the increment key is dep_date, a dimension that is defined by the departure column in the PDT, which is an alias for the flights.leaving_time column in the database. Therefore, the filter points to flights.leaving_time:
WHERE {% incrementcondition %} flights.leaving_time {%  endincrementcondition %}

You can add to the WHERE clause to create other filters. For example, if the database table goes back many years, you can create a filter so that the initial build of the PDT uses only data after a certain date. This WHERE creates a PDT with data from after January 1, 2020:

WHERE {% incrementcondition %} flights.leaving_time {%  endincrementcondition %}
  AND flights.leaving_time > '2020-01-01'

You can also use the WHERE clause to parse data in SQL into a timestamp and then give it an alias. For example, the following incremental PDT uses a 15-minute increment that is based on the text_column, which is string data that has been parsed into timestamp data:

view: sql_based_incremental_15min_pdt {
  derived_table: {
    datagroup_trigger: flights_default_datagroup
    increment_key: "event_minute15"
    increment_offset: 1
    sql: SELECT PARSE_TIMESTAMP("%c", flights.text_column) as parsed_timestamp_column,
        flights.id2  AS "id",
        flights.origin  AS "origin",
      FROM public.flights  AS flights
      WHERE {% incrementcondition %} PARSE_TIMESTAMP("%c", flights.text_column)
          {% endincrementcondition %} ;;
  }

  dimension_group: event {
    type: time
    timeframes: [raw, minute15, hour, date, week, month, year]
    datatype: timestamp
    sql:  ${TABLE}.parsed_timestamp_column ;;
  }
  dimension: id {
    type: number
  }
  dimension: origin {
    type: string
  }
}

You can use the alias for the SQL in the dimension group sql definition, but you must use the SQL expression in the WHERE clause. And then, because minute15 was set up as timeframe in the event dimension group, you can then use event_minute15 as the increment key to get a 15-minute increment for the PDT.

Creating an incremental aggregate table

To make an incremental aggregate table, add increment_key and (optionally) increment_offset under the materialization parameter of the aggregate_table parameter. Use the increment_key parameter to specify the name of a time-based LookML dimension. The dimension must be defined in the view on which the aggregate table's Explore is based.

For example, this aggregate table is based on the accidents Explore, which in this case is based on the accidents view. The aggregate table is rebuilt in increments of one week (increment_key: event_week), going back two weeks (increment_offset: 2):

explore: accidents {
  . . .
  aggregate_table: accidents_daily {
    query: {
      dimensions: [event_date, id, weather_condition]
      measures: [count]
    }
    materialization: {
      datagroup_trigger: flights_default_datagroup
      increment_key: "event_week"
      increment_offset: 2
    }
  }
}

The increment key uses the event_week dimension, which is based on the week timeframe from the event dimension group. (See the dimension_group parameter documentation page for an overview of how dimension groups work.) The dimension group and timeframe are both defined in the accidents view:

. . .
view: accidents {
  . . .
  dimension_group: event {
      type: time
      timeframes: [
        raw,
        date,
        week,
        year
      ]
      sql: ${TABLE}.event_date ;;
  }
  . . .
}

Things to consider

Optimize the source table for time-based queries

Make sure that the incremental PDT's source table is optimized for time-based queries. Specifically, the time-based column used for the increment key must have an optimization strategy, such as partitioning, sortkeys, indexes, or whatever optimization strategy is supported for your dialect. Source table optimization is strongly recommended because each time the incremental table is updated, Looker queries the source table to determine the latest values of the time-based column used for the increment key. If the source table is not optimized for these queries, Looker's query for the latest values may be slow and expensive.

Supported database dialects for incremental PDTs

For Looker to support incremental PDTs in your Looker project, your database dialect must support Data Definition Language (DDL) commands that enable deleting and inserting rows.

The following table shows which dialects support incremental PDTs in the latest release of Looker:

Dialect Supported?
Actian Avalanche
No
Amazon Athena
No
Amazon Aurora MySQL
No
Amazon Redshift
Yes
Apache Druid
No
Apache Druid 0.13+
No
Apache Druid 0.18+
No
Apache Hive 2.3+
No
Apache Hive 3.1.2+
No
Apache Spark 3+
No
ClickHouse
No
Cloudera Impala 3.1+
No
Cloudera Impala 3.1+ with Native Driver
No
Cloudera Impala with Native Driver
No
DataVirtuality
No
Databricks
Yes
Denodo 7
No
Denodo 8
No
Dremio
No
Dremio 11+
No
Exasol
No
Firebolt
No
Google BigQuery Legacy SQL
No
Google BigQuery Standard SQL
Yes
Google Cloud PostgreSQL
Yes
Google Cloud SQL
No
Google Spanner
No
Greenplum
Yes
HyperSQL
No
IBM Netezza
No
MariaDB
No
Microsoft Azure PostgreSQL
Yes
Microsoft Azure SQL Database
No
Microsoft Azure Synapse Analytics
Yes
Microsoft SQL Server 2008+
No
Microsoft SQL Server 2012+
No
Microsoft SQL Server 2016
No
Microsoft SQL Server 2017+
No
MongoBI
No
MySQL
Yes
MySQL 8.0.12+
Yes
Oracle
No
Oracle ADWC
No
PostgreSQL 9.5+
Yes
PostgreSQL pre-9.5
Yes
PrestoDB
No
PrestoSQL
No
SAP HANA 2+
No
SingleStore
No
SingleStore 7+
No
Snowflake
Yes
Teradata
No
Trino
No
Vector
No
Vertica
Yes