Filters

A filter represents a subset of records within a collection. As such, a filter always exists in the context of a given collection.

It is used to restrict the records which will be targeted by specific actions (list, update, aggregate, ...).

{
  // Condition tree
  "conditionTree": { "field": "createdAt", "operator": "today" },

  // Paging
  "page": { "limit": 30, "skip": 0 },
  "sort": [
    { "field": "title", "ascending": true },
    { "field": "id", "ascending": true }
  ],

  // Search [advanced]
  "search": "John Smith",
  "searchExtended": false,

  // Segment [advanced]
  "segment": "Active Records"
}

This is quite complex! Let's break it down into small understandable pieces.

Condition trees

A condition tree, as its names imply, is a set of conditions that applies to the records themselves and tells which records should be included or excluded from a given query.

Examples

Simple condition tree

{ "field": "title", "operator": "StartsWith", "value": "Found" }

With multiple conditions

{
  "aggregator": "And",
  "conditions": [
    { "field": "title", "operator": "Equal", "value": "Foundation" },
    { "field": "subTitle", "operator": "Equal", "value": "The Psychohistorians" }
  ]
}

With relationships

one-to-one and many-to-one relationships are supported.

In this example, we want to apply a condition tree to a related field value.

{ "field": "book:title", "operator": "Equal", "value": "Foundation" }

Of course, you can chain as many relations as you like.

{ "field": "book:price:value", "operator": "Equal", "value": 15 }

Structure

Each node of a condition tree can be one of five things:

  • An "And" branch: { aggregator: 'And', conditions: [<otherNodes>] }

  • An "Or" branch: { aggregator: 'Or', conditions: [<otherNodes>] }

  • A "condition" leaf without parameter: { field: 'title', operator: 'Present' }

  • A "condition" leaf with parameter: { field: 'title', operator: 'Equal', value: 'Foundation' }

Operators

Here is the list of operators which are supported by forest admin.

Operator
Types
Expected parameter

Present

All

Blank

All

Missing

All

Equal

All but array

Field type

NotEqual

All but array

Field type

LessThan

All but array

Field type

GreaterThan

All but array

Field type

In

All but array

Array of field type

NotIn

All but array

Array of field type

Matches

String

String

ILike

String

String

Like

String

String

StartsWith

String

String

EndsWith

String

String

Contains

String

String

NotContains

String

String

LongerThan

String

Number

ShorterThan

String

Number

Before

Date

Date

After

Date

Date

AfterXHoursAgo

Date

Number

BeforeXHoursAgo

Date

Number

Past

Date

Future

Date

PreviousMonthToDate

Date

PreviousMonth

Date

PreviousQuarterToDate

Date

PreviousQuarter

Date

PreviousWeekToDate

Date

PreviousWeek

Date

PreviousXDaysToDate

Date

Number

PreviousXDays

Date

Number

PreviousYearToDate

Date

PreviousYear

Date

Today

Date

Yesterday

Date

IncludesAll

Array

Array

Operator equivalence

You may have noticed that many operators overlap. To make data sources quicker to implement, Forest Admin supports automatic operator replacement.

What that means is that when an operator can be expressed using a combination of other operators, Forest Admin will perform the substitution automatically using the following table.

Operator
Automatic replacement

Present

NotEqual null and NotEqual ""

Blank

Equal null or Equal ""

Missing

Equal null

Equal

In [$value]

NotEqual

NotIn [$value]

In

Equal $value or Equal $2 or ...

NotIn

NotEqual $value and NotEqual $2 and ...

StartsWith

Like '$value%'

EndsWith

Like '%$value'

Contains

Like '%$value%'

Before

LessThan $value

After

GreaterThan $value

AfterXHoursAgo

GreaterThan $hoursAgo($value)

BeforeXHoursAgo

LessThan $hoursAgo($value)

Past

LessThan $now

Future

GreaterThan $now

PreviousMonthToDate

GreaterThan $startOfMonth & LessThan $now

PreviousMonth

GreaterThan $startOfLastMonth & LessThan $endOfLastMonth

PreviousQuarterToDate

GreaterThan $startOfQuarter & LessThan $now

PreviousQuarter

GreaterThan $startOfLastQuarter & LessThan $endOfLastQuarter

PreviousWeekToDate

GreaterThan $startOfWeek & LessThan $now

PreviousWeek

GreaterThan $startOfLastWeek & LessThan $endOfLastWeek

PreviousXDaysToDate

GreaterThan $xDaysAgo($value) & LessThan $now

PreviousXDays

GreaterThan $xDaysAgo($value) & LessThan $startOfToday

PreviousYearToDate

GreaterThan $startOfYear & LessThan $now

PreviousYear

GreaterThan $startOfLastYear & LessThan $endOfLastYear

Today

GreaterThan $startOfToday and LessThan $endOfToday

Yesterday

GreaterThan $startOfYesterday and LessThan $endOfYesterday

In practice:

  • if a field supports Equal, it will automatically support Blank, Missing, and In

  • if a field supports LessThan, it will automatically support Before, BeforeXHoursAgo and Past

  • ... and so on

The minimal list of operators which is sufficient to have them all is the following:

  • In and NotIn (unlocks Present, Blank, Missing, Equal, and NotEqual)

  • LessThan and GreaterThan (unlocks all dates operators)

  • Like (unlocks StartsWith, EndsWith, and Contains)

  • NotContains, LongerThan, ShorterThan, and IncludesAll

Paging

A paging clause tells the data source which page of the data should be retrieved.

Examples

{
  "page": { "limit": 30, "skip": 0 },
  "sort": [
    { "field": "title", "ascending": true },
    { "field": "id", "ascending": true }
  ]
}

The search field is a simple filter that the final user types in the search bar in the admin panel, and can be used to restrict records.

Likewise searchExtended boolean is an action that can be triggered by end-users and its implementation can vary between data sources.

For instance, in @forestadmin/datasource-sql, the searchExtended flag is used to also search content into all collections which are linked with a many-to-one or one-to-one relation to the current one.

Examples

Search into the current collection:

{ "search": "Isaac", "searchExtended": false }

Search into current and linked collections:

{ "search": "Isaac", "searchExtended": true }

Segments

The segment field in a filter contains the name of the segment which is being targeted.

Examples

{ "segment": "Active records" }

Last updated

Revision created

docs(datasource-sql): document introspection process