Struct PointFields 

pub struct PointFields(/* private fields */);

Record properties read with one point read per typed definition instead of rebuilding the record (READ-1: sorting 667K matches, or checking an unindexed age < 25 on them, rebuilt every one). Covers scalar record properties only; Self::resolve answers None for a system field or a property stored as an array anywhere, and the caller rebuilds the record.

A record holds one value per property: a write deletes the value of every typed definition of the property before it stores the new one (also when one batch writes the property twice). So the first definition holding a value is the only one, and definitions are probed in the order most records hold them: by descending owner support (under the read’s labels). A CSV column inferred as both Int64 and Float64 is read with one point read per row for most rows, not two.

Implementations

impl PointFields

pub fn resolve<'n>( snap: &QuerySnapshot, tenant: TenantId, names: impl IntoIterator<Item = &'n str>, ) -> Result<Option<Self>, QueryError>

Point reads for names, each row keyed by the name as given; None when one of them needs the whole record. Definitions are probed by their tenant-wide support.

Errors

Returns storage read errors from the property catalog.

pub fn resolve_in<'n>( snap: &QuerySnapshot, tenant: TenantId, labels: &[String], names: impl IntoIterator<Item = &'n str>, ) -> Result<Option<Self>, QueryError>

Self::resolve for records of labels (every label when empty): definitions are probed by their support under those labels.

Errors

Returns storage read errors from the property catalog.

pub fn row( &self, snap: &QuerySnapshot, tenant: TenantId, ord: u64, ) -> Result<Map<String, Json>, QueryError>

The fields of record ord: each one’s value, as the rebuilt record would carry it.

Errors

Returns storage read errors and corrupt-occurrence errors.