Struct ProfileCollector 

pub struct ProfileCollector {
    pub operators: HashMap<OperatorProfileKey, OperatorProfile>,
    pub total_rows: u64,
    pub total_time_us: u64,
    pub start_time: Option<Instant>,
    pub end_time: Option<Instant>,
    pub parallel_wall_time_us: u64,
    pub parallel_work_time_us: u64,
    pub parallel_workers: usize,
    pub parallel_buffered_chunks_peak: usize,
    pub parallel_buffered_bytes_peak: usize,
}

Collects execution profile data across all operators (for EXPLAIN output).

Fields

operators: HashMap<OperatorProfileKey, OperatorProfile>

total_rows: u64

total_time_us: u64

start_time: Option<Instant>

end_time: Option<Instant>

parallel_wall_time_us: u64

Wall-clock time spent in parallel partition execution.

parallel_work_time_us: u64

Sum of per-worker execution time (may exceed wall time).

parallel_workers: usize

Maximum number of workers used by any coordinator in this query.

parallel_buffered_chunks_peak: usize

Peak number of chunks retained in output queues.

parallel_buffered_bytes_peak: usize

Peak accounted bytes retained in output queues.

Implementations

impl ProfileCollector

pub fn new() -> Self

pub fn record_start(&mut self)

pub fn record_end(&mut self)

pub fn add_rows(&mut self, count: u64)

pub fn record_operator_profile(&mut self, profile: OperatorProfile)

pub fn parallel_profile(&self) -> (u64, u64, usize, usize, usize)

Return a snapshot of the parallel profile fields for EXPLAIN / PROFILE output.

pub fn aggregate_partition_profiles( &mut self, partition_profiles: &[ProfileCollector], )

Aggregate profiles from partition execution into this collector.

For each operator node_id, sums timing/output_rows and takes the max of peak_memory_bytes across partitions.

Trait Implementations

impl Debug for ProfileCollector

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

Formats the value using the given formatter.

impl Default for ProfileCollector

fn default() -> Self

Returns the “default value” for a type.