Operators

Vincent Depassier · August 29, 2026

PraxQL — Operators

Operators appear in where conditions, and in having for aggregation queries. Every condition has the form:

{ "field": "ColumnName", "op": "operator", "value": someValue }

Operator names are matched case-insensitively: isNull, isnull and ISNULL are the same operator.

Comparison

Operator	SQL	Description
`eq`	`=`	Equals
`neq`	`!=`	Not equals
`gt`	`>`	Greater than
`gte`	`>=`	Greater than or equal
`lt`	`<`	Less than
`lte`	`<=`	Less than or equal

{ "field": "Total", "op": "gte", "value": 100 }
{ "field": "Status", "op": "neq", "value": "Deleted" }
{ "field": "CreatedDate", "op": "gt", "value": "2026-01-01" }

Pattern matching

Operator	SQL	Description
`like`	`LIKE`	Case-sensitive pattern. `%` is the wildcard.
`ilike`	`ILIKE`	Case-insensitive pattern. `%` is the wildcard.

{ "field": "Name", "op": "like", "value": "%Corp%" }
{ "field": "Email", "op": "ilike", "value": "%@gmail.com" }

A pattern longer than 200 characters is rejected.

Set membership

Operator	SQL	Description
`in`	`IN (...)`	The value is one of the array
`notIn`	`NOT IN (...)`	The value is none of them. `nin` is accepted as an alias.

{ "field": "Status", "op": "in", "value": ["Active", "Pending", "Trial"] }
{ "field": "Status", "op": "notIn", "value": ["Deleted", "Archived"] }

value must be an array.

Null checks

Operator	SQL	Description
`isNull`	`IS NULL`	The column has no value
`isNotNull`	`IS NOT NULL`	The column has a value
`is`	`IS NULL` / `IS NOT NULL`	The same test, decided by whether you send a value

{ "field": "DeletedAt", "op": "isNull" }
{ "field": "DeletedAt", "op": "isNotNull" }

Prefer isNull and isNotNull. They say outright what is expresses through the presence or absence of a value, which is the part nobody guesses: with is, sending "value": null tests IS NULL, and omitting the key entirely tests IS NOT NULL.

To test a boolean column, compare it instead:

{ "field": "IsVerified", "op": "eq", "value": true }

Send a real JSON true, never the string "true". A string matches zero rows and reports no error, which makes it one of the slowest bugs to notice.

Range

Operator	SQL	Description
`between`	`BETWEEN x AND y`	Inclusive range

{ "field": "Date", "op": "between", "value": ["2026-01-01", "2026-06-30"] }
{ "field": "Score", "op": "between", "value": [70, 100] }

value must be a two-element array, [min, max]. Both bounds are inclusive.

Array and text

Operator	Description
`contains`	An array column contains all the given values
`textsearch`	Full-text search on a text column

{ "field": "Tags", "op": "contains", "value": ["urgent", "billing"] }
{ "field": "Description", "op": "textsearch", "value": "warehouse logistics" }

textsearch suits long-text columns. It returns results unranked — there is no relevance score, so you cannot sort by "best match".

Logical grouping

AND, the default

Conditions in the same array are AND'd automatically:

"where": [
  { "field": "IsActive", "op": "eq", "value": true },
  { "field": "Region",   "op": "eq", "value": "EU" }
]

→ WHERE IsActive = true AND Region = 'EU'

OR

"where": [
  {
    "or": [
      { "field": "Region", "op": "eq", "value": "EU" },
      { "field": "Region", "op": "eq", "value": "US" }
    ]
  }
]

→ WHERE (Region = 'EU' OR Region = 'US')

Nesting

"where": [
  { "field": "IsActive", "op": "eq", "value": true },
  {
    "or": [
      { "field": "Region", "op": "eq", "value": "EU" },
      {
        "and": [
          { "field": "Region", "op": "eq", "value": "US" },
          { "field": "Tier",   "op": "eq", "value": "Enterprise" }
        ]
      }
    ]
  }
]

→ WHERE IsActive = true AND (Region = 'EU' OR (Region = 'US' AND Tier = 'Enterprise'))

Guardrails

Limit	Value
Where conditions per query	50
Condition nesting depth	5
`like` / `ilike` pattern length	200 characters

Exceeding any of them returns 400 INVALID_QUERY rather than a truncated result. If you hit the nesting cap, the query is usually asking two questions at once — split it.