Enum Filter 

#[repr(C)]
pub enum Filter {
    None,
    Equals(FieldName, FilterValue),
    NotEquals(FieldName, FilterValue),
    Lt(FieldName, FilterValue),
    Lte(FieldName, FilterValue),
    Gt(FieldName, FilterValue),
    Gte(FieldName, FilterValue),
    In(FieldName, ValueList),
    NotIn(FieldName, ValueList),
    Contains(FieldName, FilterValue),
    StartsWith(FieldName, FilterValue),
    EndsWith(FieldName, FilterValue),
    IsNull(FieldName),
    IsNotNull(FieldName),
    And(Box<[Filter]>),
    Or(Box<[Filter]>),
    Not(Box<Filter>),
}

A complete filter that can be converted to SQL.

Size Optimization

The Filter enum is designed to fit in a single cache line (~64 bytes):

• Field names use Cow<'static, str> (24 bytes)
• Filter values use FilterValue (40 bytes)
• IN/NOT IN use Vec<FilterValue> (24 bytes) instead of SmallVec
• AND/OR use Box<[Filter]> (16 bytes)

This enables efficient iteration and better CPU cache utilization.

Zero-Allocation Patterns

For maximum performance, use static strings:

use prax_query::filter::{Filter, FilterValue};
// Zero allocation - static string borrowed
let filter = Filter::Equals("id".into(), FilterValue::Int(42));

Variants

None

No filter (always true).

Equals(FieldName, FilterValue)

Equals comparison.

NotEquals(FieldName, FilterValue)

Not equals comparison.

Lt(FieldName, FilterValue)

Less than comparison.

Lte(FieldName, FilterValue)

Less than or equal comparison.

Gt(FieldName, FilterValue)

Greater than comparison.

Gte(FieldName, FilterValue)

Greater than or equal comparison.

In(FieldName, ValueList)

In a list of values.

NotIn(FieldName, ValueList)

Not in a list of values.

Contains(FieldName, FilterValue)

Contains (LIKE %value%).

StartsWith(FieldName, FilterValue)

Starts with (LIKE value%).

EndsWith(FieldName, FilterValue)

Ends with (LIKE %value).

IsNull(FieldName)

Is null check.

IsNotNull(FieldName)

Is not null check.

And(Box<[Filter]>)

Logical AND of multiple filters.

Uses Box<[Filter]> instead of Vec<Filter> to save 8 bytes per filter (no capacity field needed since filters are immutable after construction).

Or(Box<[Filter]>)

Logical OR of multiple filters.

Uses Box<[Filter]> instead of Vec<Filter> to save 8 bytes per filter (no capacity field needed since filters are immutable after construction).

Not(Box<Filter>)

Logical NOT of a filter.

Implementations

impl Filter

pub fn none() -> Self

Create an empty filter (matches everything).

pub fn is_none(&self) -> bool

Check if this filter is empty.

pub fn and(filters: impl IntoIterator<Item = Filter>) -> Self

Create an AND filter from an iterator of filters.

Automatically filters out None filters and simplifies single-element combinations.

For known small counts, prefer and2, and3, and5, or and_n for better performance.

pub fn and2(a: Filter, b: Filter) -> Self

Create an AND filter from exactly two filters.

More efficient than and([a, b]) - avoids Vec allocation.

pub fn or(filters: impl IntoIterator<Item = Filter>) -> Self

Create an OR filter from an iterator of filters.

Automatically filters out None filters and simplifies single-element combinations.

For known small counts, prefer or2, or3, or or_n for better performance.

pub fn or2(a: Filter, b: Filter) -> Self

Create an OR filter from exactly two filters.

More efficient than or([a, b]) - avoids Vec allocation.

pub fn and_n<const N: usize>(filters: [Filter; N]) -> Self

Create an AND filter from a fixed-size array (const generic).

This is the most efficient way to create AND filters when the count is known at compile time. It avoids Vec allocation entirely.

Examples

use prax_query::filter::{Filter, FilterValue};

let filter = Filter::and_n([
    Filter::Equals("a".into(), FilterValue::Int(1)),
    Filter::Equals("b".into(), FilterValue::Int(2)),
    Filter::Equals("c".into(), FilterValue::Int(3)),
]);

pub fn or_n<const N: usize>(filters: [Filter; N]) -> Self

Create an OR filter from a fixed-size array (const generic).

This is the most efficient way to create OR filters when the count is known at compile time. It avoids Vec allocation entirely.

pub fn and3(a: Filter, b: Filter, c: Filter) -> Self

Create an AND filter from exactly 3 filters.

pub fn and4(a: Filter, b: Filter, c: Filter, d: Filter) -> Self

Create an AND filter from exactly 4 filters.

pub fn and5(a: Filter, b: Filter, c: Filter, d: Filter, e: Filter) -> Self

Create an AND filter from exactly 5 filters.

pub fn or3(a: Filter, b: Filter, c: Filter) -> Self

Create an OR filter from exactly 3 filters.

pub fn or4(a: Filter, b: Filter, c: Filter, d: Filter) -> Self

Create an OR filter from exactly 4 filters.

pub fn or5(a: Filter, b: Filter, c: Filter, d: Filter, e: Filter) -> Self

Create an OR filter from exactly 5 filters.

pub fn in_i64( field: impl Into<FieldName>, values: impl IntoIterator<Item = i64>, ) -> Self

Create an IN filter from an iterator of i64 values.

This is optimized for integer lists, avoiding the generic Into<FilterValue> conversion overhead.

pub fn in_i32( field: impl Into<FieldName>, values: impl IntoIterator<Item = i32>, ) -> Self

Create an IN filter from an iterator of i32 values.

pub fn in_strings( field: impl Into<FieldName>, values: impl IntoIterator<Item = String>, ) -> Self

Create an IN filter from an iterator of string values.

pub fn in_values(field: impl Into<FieldName>, values: ValueList) -> Self

Create an IN filter from a pre-built ValueList.

Use this when you’ve already constructed a ValueList to avoid re-collection.

pub fn in_range(field: impl Into<FieldName>, range: Range<i64>) -> Self

Create an IN filter from a range of i64 values.

Highly optimized for sequential integer ranges.

pub fn in_i64_slice(field: impl Into<FieldName>, values: &[i64]) -> Self

Create an IN filter from a pre-allocated i64 slice with exact capacity.

This is the most efficient way to create IN filters for i64 values when you have a slice available.

pub fn in_i32_slice(field: impl Into<FieldName>, values: &[i32]) -> Self

Create an IN filter for i32 values from a slice.

pub fn in_str_slice(field: impl Into<FieldName>, values: &[&str]) -> Self

Create an IN filter for string values from a slice.

pub fn not(filter: Filter) -> Self

Create a NOT filter.

pub fn in_slice<T: Into<FilterValue> + Clone>( field: impl Into<FieldName>, values: &[T], ) -> Self

Create an IN filter from a slice of values.

This is more efficient than Filter::In(field, values.into()) when you have a slice, as it avoids intermediate collection.

Examples

use prax_query::filter::Filter;

let ids: &[i64] = &[1, 2, 3, 4, 5];
let filter = Filter::in_slice("id", ids);

pub fn not_in_slice<T: Into<FilterValue> + Clone>( field: impl Into<FieldName>, values: &[T], ) -> Self

Create a NOT IN filter from a slice of values.

Examples

use prax_query::filter::Filter;

let ids: &[i64] = &[1, 2, 3, 4, 5];
let filter = Filter::not_in_slice("id", ids);

pub fn in_array<T: Into<FilterValue>, const N: usize>( field: impl Into<FieldName>, values: [T; N], ) -> Self

Create an IN filter from an array (const generic).

This is useful when you know the size at compile time.

Examples

use prax_query::filter::Filter;

let filter = Filter::in_array("status", ["active", "pending", "processing"]);

pub fn not_in_array<T: Into<FilterValue>, const N: usize>( field: impl Into<FieldName>, values: [T; N], ) -> Self

Create a NOT IN filter from an array (const generic).

pub fn and_then(self, other: Filter) -> Self

Combine with another filter using AND.

pub fn or_else(self, other: Filter) -> Self

Combine with another filter using OR.

pub fn to_sql(&self, param_offset: usize) -> (String, Vec<FilterValue>)

Generate SQL for this filter with parameter placeholders. Returns (sql, params) where params are the values to bind.

pub fn and_builder(capacity: usize) -> AndFilterBuilder

Create a builder for constructing AND filters with pre-allocated capacity.

This is more efficient than using Filter::and() when you know the approximate number of conditions upfront.

Examples

use prax_query::filter::{Filter, FilterValue};

// Build an AND filter with pre-allocated capacity for 3 conditions
let filter = Filter::and_builder(3)
    .push(Filter::Equals("active".into(), FilterValue::Bool(true)))
    .push(Filter::Gt("score".into(), FilterValue::Int(100)))
    .push(Filter::IsNotNull("email".into()))
    .build();

pub fn or_builder(capacity: usize) -> OrFilterBuilder

Create a builder for constructing OR filters with pre-allocated capacity.

This is more efficient than using Filter::or() when you know the approximate number of conditions upfront.

Examples

use prax_query::filter::{Filter, FilterValue};

// Build an OR filter with pre-allocated capacity for 2 conditions
let filter = Filter::or_builder(2)
    .push(Filter::Equals("role".into(), FilterValue::String("admin".into())))
    .push(Filter::Equals("role".into(), FilterValue::String("moderator".into())))
    .build();

pub fn builder() -> FluentFilterBuilder

Create a general-purpose filter builder.

Use this for building complex filter trees with a fluent API.

Examples

use prax_query::filter::Filter;

let filter = Filter::builder()
    .eq("status", "active")
    .gt("age", 18)
    .is_not_null("email")
    .build_and();

Trait Implementations

impl Clone for Filter

fn clone(&self) -> Filter

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for Filter

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for Filter

fn default() -> Filter

Returns the “default value” for a type.

impl From<StaticFilter> for Filter

fn from(f: StaticFilter) -> Self

Converts to this type from the input type.

impl IntoFilter for Filter

fn into_filter(self) -> Filter

Convert this type into a filter.

impl PartialEq for Filter

fn eq(&self, other: &Filter) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl StructuralPartialEq for Filter