Documentation

Define custom partitions

Use the influxctl CLI to define custom partition strategies when creating a database or table. By default, InfluxDB Cloud Dedicated partitions data by day.

The partitioning strategy of a database or table is determined by a partition template which defines the naming pattern for partition keys. Partition keys uniquely identify each partition. When a partition template is applied to a database, it becomes the default template for all tables in that database, but can be overridden when creating a table.

Partition templates can only be applied on create

You can only apply a partition template when creating a database or table. There is no way to update a partition template on an existing resource.

Use the following command flags to identify partition template parts:

  • --template-tag: An InfluxDB tag to use in the partition template.
  • --template-tag-bucket: An InfluxDB tag and number of “buckets” to group tag values into. Provide the tag key and the number of buckets to bucket tag values into separated by a comma: tagKey,N.
  • --template-timeformat: A Rust strftime date and time string that specifies the time format in the partition template and determines the time interval to partition by.

A partition template can include up to 7 total tag and tag bucket parts and only 1 time part.

View partition template part restrictions.

Always provide a time format when using custom partitioning

When defining a custom partition template for your database or table using any of the influxctl --template-* flags, always include the --template-timeformat flag with a time format to use in your partition template. Otherwise, InfluxDB omits time from the partition template and won’t compact partitions.

Create a database with a custom partition template

The following example creates a new example-db database and applies a partition template that partitions by distinct values of two tags (room and sensor-type), bucketed values of the customerID tag, and by week using the time format %Y wk:%W:

influxctl database create \
  --template-tag room \
  --template-tag sensor-type \
  --template-tag-bucket customerID,500 \
  --template-timeformat '%Y wk:%W' \
  example-db

Create a table with a custom partition template

The following example creates a new example-table table in the example-db database and applies a partition template that partitions by distinct values of two tags (room and sensor-type), bucketed values of the customerID tag, and by month using the time format %Y-%m:

influxctl table create \
  --template-tag room \
  --template-tag sensor-type \
  --template-tag-bucket customerID,500 \
  --template-timeformat '%Y-%m' \
  example-db \
  example-table

Example partition templates

Given the following line protocol with a 2024-01-01T00:00:00Z timestamp:

prod,line=A,station=weld1 temp=81.9,qty=36i 1704067200000000000
Partitioning by distinct tag values
DescriptionTag part(s)Time partResulting partition key
By day (default)%Y-%m-%d2024-01-01
By day (non-default)%d %b %Y01 Jan 2024
By week%Y wk:%W2024 wk:01
By month%Y-%m2024-01
Single tag, by dayline%FA | 2024-01-01
Single tag, by weekline%Y wk:%WA | 2024 wk:01
Single tag, by monthline%Y-%mA | 2024-01
Multiple tags, by dayline, station%FA | weld1 | 2024-01-01
Multiple tags, by weekline, station%Y wk:%WA | weld1 | 2024 wk:01
Multiple tags, by monthline, station%Y-%mA | weld1 | 2024-01
Partition by tag buckets
DescriptionTag partTag bucket partTime partResulting partition key
Distinct tag, tag buckets, by daylinestation,100%FA | 3 | 2024-01-01
Distinct tag, tag buckets, by weeklinestation,500%Y wk:%WA | 303 | 2024 wk:01

Was this page helpful?

Thank you for your feedback!


The future of Flux

Flux is going into maintenance mode. You can continue using it as you currently are without any changes to your code.

Read more