Skip to content
Query.Farm
Talk with Us

Filter pushdown

On this page

Receiving and evaluating pushed-down WHERE predicates.

struct AndFilter

Section titled “struct AndFilter”
source
type AndFilter struct {
columnName  string
columnIndex  int
Children  []Filter
}

Description

AndFilter is a conjunction of child filters. All children must pass.

Methods

method ColumnIndex

Section titled “method ColumnIndex”
source
func (f *AndFilter) ColumnIndex() int

ColumnIndex returns the index of the column this filter applies to.

method ColumnName

Section titled “method ColumnName”
source
func (f *AndFilter) ColumnName() string

ColumnName returns the name of the column this filter applies to.

method Evaluate

Section titled “method Evaluate”
source
func (f *AndFilter) Evaluate(ctx context.Context, batch arrow.RecordBatch) (arrow.Array, error)

Evaluate combines its child filters with a Kleene logical AND, returning a boolean array that is true only where every child passes.

method Type

Section titled “method Type”
source
func (f *AndFilter) Type() FilterType

Type returns the filter type identifier, FilterAnd.

struct ColumnBounds

Section titled “struct ColumnBounds”
source
type ColumnBounds struct {
// MinValue is the minimum bound value, or nil if unbounded below.
MinValue  scalar.Scalar
// MinInclusive is true if MinValue is inclusive (>=), false if exclusive (>).
MinInclusive  bool
// MaxValue is the maximum bound value, or nil if unbounded above.
MaxValue  scalar.Scalar
// MaxInclusive is true if MaxValue is inclusive (<=), false if exclusive (<).
MaxInclusive  bool
}

Description

ColumnBounds represents numeric/comparable bounds for a column extracted from comparison filters.

type ComparisonOp

Section titled “type ComparisonOp”
source
type ComparisonOp string

Description

ComparisonOp identifies a comparison operator used in constant filters.

Methods

method Symbol

Section titled “method Symbol”
source
func (op ComparisonOp) Symbol() string

Symbol returns the SQL operator string for this comparison op.

method computeFuncName

Section titled “method computeFuncName”
source
func (op ComparisonOp) computeFuncName() string

computeFuncName returns the arrow compute function name for this operator.

struct ConstantFilter

Section titled “struct ConstantFilter”
source
type ConstantFilter struct {
columnName  string
columnIndex  int
Op    ComparisonOp
Value    scalar.Scalar
}

Description

ConstantFilter compares a column against a constant value.

Methods

method ColumnIndex

Section titled “method ColumnIndex”
source
func (f *ConstantFilter) ColumnIndex() int

ColumnIndex returns the index of the column this filter applies to.

method ColumnName

Section titled “method ColumnName”
source
func (f *ConstantFilter) ColumnName() string

ColumnName returns the name of the column this filter applies to.

method Evaluate

Section titled “method Evaluate”
source
func (f *ConstantFilter) Evaluate(ctx context.Context, batch arrow.RecordBatch) (arrow.Array, error)

Evaluate compares the column against the constant value using the configured comparison operator, returning a boolean array of matches.

method Type

Section titled “method Type”
source
func (f *ConstantFilter) Type() FilterType

Type returns the filter type identifier, FilterConstant.

struct ExpressionFilter

Section titled “struct ExpressionFilter”
source
type ExpressionFilter struct {
columnName  string
columnIndex  int
Expr    *exprNode
// Values holds the scalars referenced by constant-node value_ref indices.
// Populated at parse time from the filter batch's value-ref columns.
Values  []scalarValueRef
}

Description

ExpressionFilter is a recursive expression tree pushed from DuckDB. The worker evaluates it via an embedded DuckDB connection (which can load the spatial extension for geometry predicates). Mirrors vgi-python’s ExpressionFilter in table_filter_pushdown.py.

Methods

method ColumnIndex

Section titled “method ColumnIndex”
source
func (f *ExpressionFilter) ColumnIndex() int

ColumnIndex returns the index of the column the filter applies to.

method ColumnName

Section titled “method ColumnName”
source
func (f *ExpressionFilter) ColumnName() string

ColumnName returns the name of the column the filter applies to.

method Evaluate

Section titled “method Evaluate”
source
func (f *ExpressionFilter) Evaluate(ctx context.Context, batch arrow.RecordBatch) (arrow.Array, error)

Evaluate renders the expression tree to SQL, feeds the input batch to a local DuckDB, and runs SELECT (<expr>)::BOOLEAN. Returns a Boolean array of length batch.NumRows().

method Type

Section titled “method Type”
source
func (f *ExpressionFilter) Type() FilterType

Type returns the filter type, which is always FilterExpression.

interface Filter

Section titled “interface Filter”
source
type Filter interface {
// ColumnName returns the name of the column this filter applies to.
ColumnName() string
// ColumnIndex returns the index of the column in the output schema.
ColumnIndex() int
// Type returns the filter type identifier.
Type() FilterType
// Evaluate evaluates the filter against a record batch, returning a
// boolean array where true indicates the row passes the filter.
Evaluate(ctx context.Context, batch arrow.RecordBatch) (arrow.Array, error)
}

Description

Filter is the interface all pushdown filter types implement.

Methods

function parseFilter

Section titled “function parseFilter”
source
func parseFilter(spec filterSpec, getValue func(int) (scalar.Scalar, error), getJoinKey func(string) arrow.Array) (Filter, error)

function parseFilterWithBatch

Section titled “function parseFilterWithBatch”
source
func parseFilterWithBatch(spec filterSpec, getValue func(int) (scalar.Scalar, error), getJoinKey func(string) arrow.Array, batch arrow.RecordBatch) (Filter, error)

parseFilterWithBatch is like parseFilter but also provides access to the raw filter batch so expression filters can harvest their value-ref columns (including Arrow extension metadata such as geoarrow.wkb).

function withColumnIndex

Section titled “function withColumnIndex”
source
func withColumnIndex(f Filter, idx int) Filter

withColumnIndex returns a new filter with the column index adjusted. This is used by StructFilter to evaluate child filters at index 0.

interface FilterPrimitive

Section titled “interface FilterPrimitive”
source
type FilterPrimitive interface {
~string | ~int64 | ~int32 | ~int16 | ~int8 |
  ~uint64 | ~uint32 | ~uint16 | ~uint8 |
  ~float64 | ~float32 | ~bool
}

Description

FilterPrimitive constrains the value type that EqualOrInValues / EqualValue can extract. Covers the scalar types DuckDB serializes through the pushdown filter wire protocol.

type FilterType

Section titled “type FilterType”
source
type FilterType string

Description

FilterType identifies the kind of pushdown filter.

struct InFilter

Section titled “struct InFilter”
source
type InFilter struct {
columnName  string
columnIndex  int
Values    arrow.Array
}

Description

InFilter checks whether a column value is in a set of values.

Methods

method ColumnIndex

Section titled “method ColumnIndex”
source
func (f *InFilter) ColumnIndex() int

ColumnIndex returns the index of the column this filter applies to.

method ColumnName

Section titled “method ColumnName”
source
func (f *InFilter) ColumnName() string

ColumnName returns the name of the column this filter applies to.

method Evaluate

Section titled “method Evaluate”
source
func (f *InFilter) Evaluate(ctx context.Context, batch arrow.RecordBatch) (arrow.Array, error)

Evaluate tests whether each column value is a member of the filter’s value set, returning a boolean array that is true for matching rows.

method Type

Section titled “method Type”
source
func (f *InFilter) Type() FilterType

Type returns the filter type identifier, FilterIn.

struct IsNotNullFilter

Section titled “struct IsNotNullFilter”
source
type IsNotNullFilter struct {
columnName  string
columnIndex  int
}

Description

IsNotNullFilter checks whether a column value is not null.

Methods

method ColumnIndex

Section titled “method ColumnIndex”
source
func (f *IsNotNullFilter) ColumnIndex() int

ColumnIndex returns the index of the column this filter applies to.

method ColumnName

Section titled “method ColumnName”
source
func (f *IsNotNullFilter) ColumnName() string

ColumnName returns the name of the column this filter applies to.

method Evaluate

Section titled “method Evaluate”
source
func (f *IsNotNullFilter) Evaluate(ctx context.Context, batch arrow.RecordBatch) (arrow.Array, error)

Evaluate tests each column value for non-null, returning a boolean array that is true where the value is not null.

method Type

Section titled “method Type”
source
func (f *IsNotNullFilter) Type() FilterType

Type returns the filter type identifier, FilterIsNotNull.

struct IsNullFilter

Section titled “struct IsNullFilter”
source
type IsNullFilter struct {
columnName  string
columnIndex  int
}

Description

IsNullFilter checks whether a column value is null.

Methods

method ColumnIndex

Section titled “method ColumnIndex”
source
func (f *IsNullFilter) ColumnIndex() int

ColumnIndex returns the index of the column this filter applies to.

method ColumnName

Section titled “method ColumnName”
source
func (f *IsNullFilter) ColumnName() string

ColumnName returns the name of the column this filter applies to.

method Evaluate

Section titled “method Evaluate”
source
func (f *IsNullFilter) Evaluate(ctx context.Context, batch arrow.RecordBatch) (arrow.Array, error)

Evaluate tests each column value for null, returning a boolean array that is true where the value is null.

method Type

Section titled “method Type”
source
func (f *IsNullFilter) Type() FilterType

Type returns the filter type identifier, FilterIsNull.

struct OrFilter

Section titled “struct OrFilter”
source
type OrFilter struct {
columnName  string
columnIndex  int
Children  []Filter
}

Description

OrFilter is a disjunction of child filters. At least one child must pass.

Methods

method ColumnIndex

Section titled “method ColumnIndex”
source
func (f *OrFilter) ColumnIndex() int

ColumnIndex returns the index of the column this filter applies to.

method ColumnName

Section titled “method ColumnName”
source
func (f *OrFilter) ColumnName() string

ColumnName returns the name of the column this filter applies to.

method Evaluate

Section titled “method Evaluate”
source
func (f *OrFilter) Evaluate(ctx context.Context, batch arrow.RecordBatch) (arrow.Array, error)

Evaluate combines its child filters with a Kleene logical OR, returning a boolean array that is true where at least one child passes.

method Type

Section titled “method Type”
source
func (f *OrFilter) Type() FilterType

Type returns the filter type identifier, FilterOr.

struct PushdownFilters

Section titled “struct PushdownFilters”
source
type PushdownFilters struct {
Filters  []Filter
Version  string
}

Description

PushdownFilters holds the deserialized pushdown filters for a function call.

Methods

method Apply

Section titled “method Apply”
source
func (pf *PushdownFilters) Apply(ctx context.Context, batch arrow.RecordBatch) (arrow.RecordBatch, error)

Apply applies all filters to the batch, returning a filtered batch.

function DeserializeFilters

Section titled “function DeserializeFilters”
source
func DeserializeFilters(batch arrow.RecordBatch, joinKeys …map[string]arrow.Array) (*PushdownFilters, error)

DeserializeFilters deserializes a pushdown filters record batch into a PushdownFilters container with a typed filter AST. Join-keys InFilters are resolved by name against joinKeys (name -> column) when provided.

method Evaluate

Section titled “method Evaluate”
source
func (pf *PushdownFilters) Evaluate(ctx context.Context, batch arrow.RecordBatch) (arrow.Array, error)

Evaluate evaluates all filters against the batch, returning a boolean mask. Filters are combined with AND at the top level.

method FilteredColumns

Section titled “method FilteredColumns”
source
func (pf *PushdownFilters) FilteredColumns() map[string]struct{}

FilteredColumns returns the set of column names that have filters applied.

method GetColumnBounds

Section titled “method GetColumnBounds”
source
func (pf *PushdownFilters) GetColumnBounds(name string) *ColumnBounds

GetColumnBounds extracts numeric bounds from comparison filters on the named column. Returns nil if no bounds can be determined.

method GetColumnConstant

Section titled “method GetColumnConstant”
source
func (pf *PushdownFilters) GetColumnConstant(name string) scalar.Scalar

GetColumnConstant returns the constant value if the column has an equality filter, or nil if no equality filter exists.

method GetColumnFilters

Section titled “method GetColumnFilters”
source
func (pf *PushdownFilters) GetColumnFilters(name string) []Filter

GetColumnFilters returns all top-level filters for a specific column.

method GetColumnInValues

Section titled “method GetColumnInValues”
source
func (pf *PushdownFilters) GetColumnInValues(name string) arrow.Array

GetColumnInValues returns the IN filter values for a column, or nil if no IN filter exists.

method GetColumnValues

Section titled “method GetColumnValues”
source
func (pf *PushdownFilters) GetColumnValues(name string) arrow.Array

GetColumnValues returns discrete values a column could have based on equality or IN filters. For EQ, wraps the value in a 1-element array. Returns nil if no discrete values can be determined.

Descends one level into AndFilter children (via collectColumnFilters), consistent with GetColumnBounds: DuckDB pushes col = v / col IN (...) conjoined with derived range bounds as a single AndFilter (e.g. a semi-join emits col IN (...) AND col >= min AND col <= max). Without the descent the discrete-value fast path silently misses those and pruning callers fall back to scanning every partition.

An OrFilter resolves to the UNION of its branches, but only when every branch pins this column to discrete values; if any branch is a range/IS NULL or constrains a different column the set is not enumerable and we return nil.

method HasFilterForColumn

Section titled “method HasFilterForColumn”
source
func (pf *PushdownFilters) HasFilterForColumn(name string) bool

HasFilterForColumn returns true if any filter constrains the given column.

method Repr

Section titled “method Repr”
source
func (pf *PushdownFilters) Repr() string

Repr returns a Python-style repr of the filter set, matching the vgi-python PushdownFilters.repr output. Useful for diagnostic display (e.g., the dynamic_filter_echo example) and stable regression testing. Format:

PushdownFilters([])                              when empty
PushdownFilters([Filter1, Filter2, ...])         otherwise

method ToSQL

Section titled “method ToSQL”
source
func (pf *PushdownFilters) ToSQL(quoteIdentifier func(string) string, placeholder string) (string, []interface{})

ToSQL converts filters to a SQL WHERE clause with parameters. The quoteIdentifier function is used to quote column names (default: double quotes). The placeholder is the parameter placeholder style (“?”, “%s”, etc.). Returns the clause (excluding “WHERE” keyword) and parameter values.

method collectColumnFilters

Section titled “method collectColumnFilters”
source
func (pf *PushdownFilters) collectColumnFilters(name string) []Filter

collectColumnFilters collects filters for a column from top-level and direct AND children.

struct StructFilter

Section titled “struct StructFilter”
source
type StructFilter struct {
columnName  string
columnIndex  int
ChildIndex  int
ChildName  string
ChildFilter  Filter
}

Description

StructFilter filters on a nested field within a struct column.

Methods

method ColumnIndex

Section titled “method ColumnIndex”
source
func (f *StructFilter) ColumnIndex() int

ColumnIndex returns the index of the column this filter applies to.

method ColumnName

Section titled “method ColumnName”
source
func (f *StructFilter) ColumnName() string

ColumnName returns the name of the column this filter applies to.

method Evaluate

Section titled “method Evaluate”
source
func (f *StructFilter) Evaluate(ctx context.Context, batch arrow.RecordBatch) (arrow.Array, error)

Evaluate applies the child filter to the nested struct field, returning the boolean array produced by evaluating that field.

method Type

Section titled “method Type”
source
func (f *StructFilter) Type() FilterType

Type returns the filter type identifier, FilterStruct.

type driver_value

Section titled “type driver_value”
source
type driver_value = driver.Value

struct exprNode

Section titled “struct exprNode”
source
type exprNode struct {
Type    string    `json:"expr_type"`
Index    int    `json:"index,omitempty"`
ValueRef  *int    `json:"value_ref,omitempty"`
FunctionName  string    `json:"function_name,omitempty"`
Children  []exprNode  `json:"children,omitempty"`
Op    string    `json:"op,omitempty"`
Left    *exprNode  `json:"left,omitempty"`
Right    *exprNode  `json:"right,omitempty"`
Conjunction  string    `json:"conjunction_type,omitempty"`
}

Methods

method toSQL

Section titled “method toSQL”
source
func (n *exprNode) toSQL(columnName string, schema *arrow.Schema, values []scalarValueRef) (string, error)

toSQL converts the expression tree to a SQL string. Column references resolve to the (double-quoted) filter column name. Constants are rendered from the scalar-values array (resolved at filter-parse time).

struct filterSpec

Section titled “struct filterSpec”
source
type filterSpec struct {
Type    string      `json:"type"`
ColumnName  string      `json:"column_name"`
ColumnIndex  int      `json:"column_index"`
Op    string      `json:"op,omitempty"`
ValueRef  *int      `json:"value_ref,omitempty"`
KeysColumn  string      `json:"keys_column,omitempty"`
Children  []filterSpec    `json:"children,omitempty"`
ChildIndex  int      `json:"child_index,omitempty"`
ChildName  string      `json:"child_name,omitempty"`
ChildFilter  *filterSpec    `json:"child_filter,omitempty"`
Expr    *json.RawMessage  `json:"expr,omitempty"`
}

struct scalarValueRef

Section titled “struct scalarValueRef”
source
type scalarValueRef struct {
value  any  // Go scalar (int64, float64, string, []byte, nil)
hex  string  // hex string for binary values
wkb  bool  // true if the value comes from a geoarrow.wkb column
}

Description

scalarValueRef captures one scalar value + its Arrow field metadata (so geoarrow.wkb constants can wrap into ST_GeomFromHEXWKB).

Methods

function resolveScalarValueRef

Section titled “function resolveScalarValueRef”
source
func resolveScalarValueRef(batch arrow.RecordBatch, ref int) (scalarValueRef, error)

resolveScalarValueRef extracts one scalar from the filter batch at column index ref+1 (column 0 holds the JSON specs).

function EqualOrInValues

Section titled “function EqualOrInValues”
source
func EqualOrInValues[T FilterPrimitive](pf *PushdownFilters, column string) ([]T, bool)

EqualOrInValues returns the values from a single column = X ConstantFilter, a single column IN (...) InFilter, or an AND-conjunction of such filters (all on the same column). Returns ok=false if the column has any other filter shape (range comparisons, IS NULL/NOT NULL, OR, nested struct, etc.) — callers should fall back to PushdownFilters.Apply for those.

The fast path eliminates the per-worker type-switch over Filter.(type) and the repeated scalar-to-Go-value conversion code that workers otherwise write to support filter pushdown on string / integer key columns.

Returns a deduplicated, order-preserving slice of values.

function EqualValue

Section titled “function EqualValue”
source
func EqualValue[T FilterPrimitive](pf *PushdownFilters, column string) (T, bool)

EqualValue is the single-value variant: returns the value of a single column = X ConstantFilter, or ok=false otherwise.

function HasOnlyEqualOrInOn

Section titled “function HasOnlyEqualOrInOn”
source
func HasOnlyEqualOrInOn(pf *PushdownFilters, columns …string) bool

HasOnlyEqualOrInOn returns true when every filter in pf is an eq/in (or AND-of-those) on one of the named columns. Useful for “either I can fast- path everything, or fall back to a slow path entirely” decisions.

function appenderForConn

Section titled “function appenderForConn”
source
func appenderForConn(ctx context.Context, conn *sql.Conn, tableName string) (*duckdb.Appender, error)

function arrayToGoSlice

Section titled “function arrayToGoSlice”
source
func arrayToGoSlice(arr arrow.Array) []interface{}

function arrowToDriverValue

Section titled “function arrowToDriverValue”
source
func arrowToDriverValue(col arrow.Array, i int, field arrow.Field) (interface{}, error)

function arrowValueAtForList

Section titled “function arrowValueAtForList”
source
func arrowValueAtForList(col arrow.Array, i int) (interface{}, error)

arrowValueAtForList extracts a Go scalar from a list’s child array.

function batchHasDictionary

Section titled “function batchHasDictionary”
source
func batchHasDictionary(batch arrow.RecordBatch) bool

batchHasDictionary reports whether any column of batch is dictionary-encoded.

function buildCreateTableSQL

Section titled “function buildCreateTableSQL”
source
func buildCreateTableSQL(name string, schema *arrow.Schema) (string, error)

function collectEqOrIn

Section titled “function collectEqOrIn”
source
func collectEqOrIn[T FilterPrimitive](f Filter, out *[]T, seen map[T]struct{}) bool

collectEqOrIn walks one filter, appending typed values to *out. Returns false if the filter (or any AND child) is not an eq/in. AND of eq/in is treated as the intersection-by-listing; callers usually want a single column filter so this composes naturally.

function collectExprValueRefs

Section titled “function collectExprValueRefs”
source
func collectExprValueRefs(n *exprNode, out map[int]bool)

collectExprValueRefs walks the expression tree collecting every value_ref index referenced by constant nodes.

function duckDBTypeFor

Section titled “function duckDBTypeFor”
source
func duckDBTypeFor(f arrow.Field) (string, error)

function ensureEvalDB

Section titled “function ensureEvalDB”
source
func ensureEvalDB() (*sql.DB, error)

function escapeIdent

Section titled “function escapeIdent”
source
func escapeIdent(s string) string

function evalExpressionAgainstBatch

Section titled “function evalExpressionAgainstBatch”
source
func evalExpressionAgainstBatch(ctx context.Context, batch arrow.RecordBatch, sqlExpr string) (arrow.Array, error)

evalExpressionAgainstBatch feeds a RecordBatch to embedded DuckDB as a temporary table, runs SELECT (<sqlExpr>)::BOOLEAN, and returns the result mask as an Arrow Boolean array.

The batch is staged via a one-row-at-a-time INSERT to avoid depending on duckdb’s Arrow integration (which requires a binary build matching the driver). This is slower than the Python version’s from_arrow but correctness-first; a future optimisation could use the appender API.

function filterArrayTo

Section titled “function filterArrayTo”
source
func filterArrayTo[T FilterPrimitive](a arrow.Array) ([]T, bool)

filterArrayTo extracts all non-null values from an Arrow array as []T. Returns ok=false on type mismatch with the target Go type.

function filterColumn

Section titled “function filterColumn”
source
func filterColumn(ctx context.Context, col, mask arrow.Array, opts compute.FilterOptions) (arrow.Array, error)

filterColumn filters a single column by mask. Dictionary columns filter their primitive index array (supported by the take kernel) and rebuild with the same dictionary values, preserving the dictionary encoding.

function filterIsEqOrInOn

Section titled “function filterIsEqOrInOn”
source
func filterIsEqOrInOn(f Filter, allowed map[string]struct{}) bool

function filterRecordBatchDictAware

Section titled “function filterRecordBatchDictAware”
source
func filterRecordBatchDictAware(ctx context.Context, batch arrow.RecordBatch, mask arrow.Array) (arrow.RecordBatch, error)

filterRecordBatchDictAware filters batch by mask column-by-column, handling dictionary columns the Go arrow fork’s array_take kernel cannot take.

function filterScalarTo

Section titled “function filterScalarTo”
source
func filterScalarTo[T FilterPrimitive](s scalar.Scalar) (T, bool)

filterScalarTo converts an Arrow scalar.Scalar to a concrete Go value of type T. Returns ok=false for type mismatches or invalid scalars.

function filterToSQL

Section titled “function filterToSQL”
source
func filterToSQL(f Filter, quote func(string) string, placeholder string) (string, []interface{})

function insertBatchToDuckDB

Section titled “function insertBatchToDuckDB”
source
func insertBatchToDuckDB(ctx context.Context, conn *sql.Conn, name string, batch arrow.RecordBatch) error

function isOperatorName

Section titled “function isOperatorName”
source
func isOperatorName(name string) bool

function isWKBField

Section titled “function isWKBField”
source
func isWKBField(f arrow.Field) bool

isWKBField returns true if the Arrow field is flagged with the geoarrow.wkb extension name.

function makeBoolArray

Section titled “function makeBoolArray”
source
func makeBoolArray(value bool, length int) arrow.Array

makeBoolArray creates a boolean array of constant value.

function nextEvalID

Section titled “function nextEvalID”
source
func nextEvalID() int64

function orDiscreteValues

Section titled “function orDiscreteValues”
source
func orDiscreteValues(or *OrFilter, name string) arrow.Array

orDiscreteValues returns the deduplicated union of discrete values for the named column across all OR branches, or nil if any branch leaves the column unbounded (a range/IS NULL branch, or a branch constraining a different column). Unlike the AND case, returning one branch’s values would be an unsafe subset — a pruning caller would skip the other branches’ rows. Descends one level only, consistent with collectColumnFilters.

function renderConstantRef

Section titled “function renderConstantRef”
source
func renderConstantRef(n *exprNode, values []scalarValueRef) (string, error)

renderConstantRef produces a SQL literal for a constant node by looking up its value_ref index in the value array.

function reprFilter

Section titled “function reprFilter”
source
func reprFilter(f Filter) string

function scalarGreater

Section titled “function scalarGreater”
source
func scalarGreater(a, b scalar.Scalar) bool

function scalarGreaterEqual

Section titled “function scalarGreaterEqual”
source
func scalarGreaterEqual(a, b scalar.Scalar) bool

function scalarLess

Section titled “function scalarLess”
source
func scalarLess(a, b scalar.Scalar) bool

function scalarLessEqual

Section titled “function scalarLessEqual”
source
func scalarLessEqual(a, b scalar.Scalar) bool

function scalarToAny

Section titled “function scalarToAny”
source
func scalarToAny(s scalar.Scalar) (any, bool)

scalarToAny extracts the underlying Go value from an Arrow scalar, returning values typed as the matching Go primitive for the FilterPrimitive set.

function scalarToArray

Section titled “function scalarToArray”
source
func scalarToArray(s scalar.Scalar) arrow.Array

function scalarToFloat64

Section titled “function scalarToFloat64”
source
func scalarToFloat64(s scalar.Scalar) (float64, bool)

function scalarToGo

Section titled “function scalarToGo”
source
func scalarToGo(s scalar.Scalar) interface{}