Filter

Definition

A filter, in Bucket, is a mechanism that is used to check if entities such as user, company, event, etc. match a set of predicates. Filters can also be aggregated into logical expressions, thus, facilitating advanced use cases.

Bucket support the following filter types:

• Company attribute filter, that can be used to check company attributes,

• Company feature metrics filter, which allows checking for feature metrics applied to a given company (for example which STARS state the company is in for a feature),

• User attribute filter, used to check user attributes,

• Event attribute filter, used to check event attributes,

• Company segment filter, can be used to check a company's membership in a segment,

• Feature access filter, can be used to check whether a company has access to a feature,

• Gradual rollout filter, is used in advanced scenarios to evaluate whether a company matches target rollout bracket,

• Other context filter, used when rules can access additional context, in addition to user and company attributes.

Company attribute filter

This filter can be used to check company attributes against a set predicates. The attributes include the First seen and Last seen which are maintained by Bucket. You can use any attribute name that your application sent to Bucket.

Company feature metrics

This filter allows checking company's feature-specific metrics. These metrics include things like Event count, First used, Last used, STARS state, end etc.

User attribute filter

This filter can be used to check user attributes against a set predicates. You can use any attribute name that your application sent to Bucket when updating user.

Event attribute filter

This filter can be used to check event attributes against a set predicates. You can use any attribute name that your application sent to Bucket when sending track events.

Company segment filter

This filter can be used to check if a given company is (or not) included in a given segment. The filter essentially evaluates the segment's filter against the company.

Feature access filter

This filter can be used to check if a given company has (or not) access to a given feature. The filter evaluates the feature's targeting rules against the provided company, and assumes skips evaluation of any non-company attribute related filters.

Gradual rollout filter

This filter is used by feature access targeting when enabling gradual rollout. It will bracket the pool of companies based on a predictable hashing algorithm and check if the company falls within the rollout percentage.

Other context filter

This filter can be used to check other context attributes against a set predicates. You can use any attribute name that your application sends to Bucket when evaluating feature access.

• Company attribute filters using First seen and Last seen attributes cannot be used in targeting rules,

• Company feature metrics filters are not supported in targeting rules,

• Event attribute filter is only used in event-based feature and automatic feedback surveys,

• Gradual rollout filter is only used in feature access targeting rules,

• Other context filter is only used in feature access targeting rules,

Any filters that build up on on other filters inherit the restrictions of the filters they are based on.

Next steps

• Learn in depth how to use filters in setup feature access rules.