Struct BlockStats 

pub struct BlockStats {
    pub min: f64,
    pub max: f64,
    pub null_count: u32,
    pub row_count: u32,
    pub str_min: Option<String>,
    pub str_max: Option<String>,
    pub bloom: Option<BloomFilter>,
    pub min_i64: Option<i64>,
    pub max_i64: Option<i64>,
}

Per-block statistics for a single column. Enables predicate pushdown: skip blocks where WHERE price > 100 and block’s max_price < 100.

Fields

min: f64

Minimum value in this block (encoded as f64 for uniformity; i64 values are cast losslessly for small values; strings use NaN).

For i64/timestamp columns, prefer min_i64 when available — f64 cannot represent all i64 values exactly (values outside ±2^53 may be rounded).

max: f64

Maximum value in this block (see min for caveats on i64 precision).

null_count: u32

Number of null values in this block.

row_count: u32

Number of rows in this block (≤ BLOCK_SIZE, last block may be smaller).

str_min: Option<String>

Lexicographic minimum for string columns (truncated to 32 bytes). None for numeric, bool, binary, vector, and other non-string columns.

str_max: Option<String>

Lexicographic maximum for string columns (truncated to 32 bytes). None for non-string columns.

bloom: Option<BloomFilter>

Bloom filter for equality-predicate skipping on string columns.

Carries the filter parameters (k, m) alongside the bytes so that readers never depend on compile-time constants to interpret the filter. None when there are no non-null string values in the block, or when cardinality is too low to justify the overhead.

min_i64: Option<i64>

Exact integer minimum for i64/timestamp columns.

Set alongside min (which holds the lossy f64 cast) so that predicates with integral values outside ±2^53 can compare losslessly. None for all non-integer column types and for segments written before minor v1.

max_i64: Option<i64>

Exact integer maximum for i64/timestamp columns (see min_i64).

Implementations

impl BlockStats

pub fn numeric(min: f64, max: f64, null_count: u32, row_count: u32) -> Self

Create stats for a numeric block with known min/max.

pub fn integer(min: i64, max: i64, null_count: u32, row_count: u32) -> Self

Create stats for an i64 or timestamp column block.

Populates both the lossless min_i64/max_i64 fields AND the lossy min/max f64 fields so that the f64 path remains a valid fallback for non-integral predicate values.

pub fn non_numeric(null_count: u32, row_count: u32) -> Self

Create stats for a block with no meaningful numeric min/max (non-numeric columns).

pub fn string_block( null_count: u32, row_count: u32, str_min: Option<String>, str_max: Option<String>, bloom: Option<BloomFilter>, ) -> Self

Create stats for a string block with lexicographic bounds and bloom filter.

Trait Implementations

impl Clone for BlockStats

fn clone(&self) -> BlockStats

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for BlockStats

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for BlockStats

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl<'__msgpack_de> FromMessagePack<'__msgpack_de> for BlockStats

fn read<R: Read<'__msgpack_de>>(reader: &mut R) -> Result<Self, Error>

where Self: Sized,

Reads the MessagePack representation of this value from the provided reader.

impl Serialize for BlockStats

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl ToMessagePack for BlockStats

fn write<W: Write>(&self, writer: &mut W) -> Result<(), Error>

Writes the MessagePack representation of this value into the provided writer.