Struct ExecutionMetadata 

pub struct ExecutionMetadata {

    pub engine: Option<String>,
    pub rows_processed: usize,
    pub bytes_consumed: Option<u64>,
    pub columns_detected: usize,
    pub scan_time_ms: u128,
    pub throughput_rows_sec: Option<f64>,
    pub memory_peak_mb: Option<f64>,
    pub error_count: usize,
    pub ragged_row_count: usize,
    pub source_exhausted: bool,
    pub truncation_reason: Option<TruncationReason>,
    pub sampling_applied: bool,
    pub sampling_ratio: Option<f64>,
}

Metadata about the profiling execution.

Fields

engine: Option<String>

Engine or parser that actually produced the report.

rows_processed: usize

Number of rows actually processed or analyzed.

bytes_consumed: Option<u64>

Number of bytes consumed from the source, if known.

columns_detected: usize

Number of columns detected in the data.

scan_time_ms: u128

Total execution time in milliseconds.

throughput_rows_sec: Option<f64>

Throughput in rows per second, auto-calculated when possible.

memory_peak_mb: Option<f64>

Peak process memory usage (resident set size) in megabytes observed during profiling, sampled at chunk/batch boundaries. Absent when the platform provided no reading or the input path does not track memory.

error_count: usize

Number of errors encountered during profiling.

ragged_row_count: usize

Number of data rows whose field count differed from the header’s.

These rows were recovered — extra trailing fields dropped, missing trailing fields padded to null — so profiling continued, but each one is a structural violation in the source. A nonzero value means the file did not parse cleanly even though a report was produced. Additive field: legacy documents without it deserialize as 0.

source_exhausted: bool

Whether the entire source was consumed.

truncation_reason: Option<TruncationReason>

If the source was not exhausted, why processing stopped.

sampling_applied: bool

Whether sampling was applied.

sampling_ratio: Option<f64>

Ratio of rows analyzed to total rows when sampling is meaningful.

Implementations

impl ExecutionMetadata

pub fn new( rows_processed: usize, columns_detected: usize, scan_time_ms: u128, ) -> Self

Create new execution metadata with throughput calculated automatically.

pub fn with_engine(self, engine: impl Into<String>) -> Self

Record the engine or parser that actually produced the report.

pub fn with_sampling(self, ratio: f64) -> Self

Set sampling information.

pub fn with_source_exhausted(self, exhausted: bool) -> Self

Explicitly set whether the source was fully consumed.

pub fn with_truncation(self, reason: TruncationReason) -> Self

Mark the execution as truncated.

pub fn with_bytes_consumed(self, bytes: u64) -> Self

Set the number of bytes consumed from the source.

pub fn with_error_count(self, count: usize) -> Self

Set the error count.

pub fn with_ragged_row_count(self, count: usize) -> Self

Set the count of ragged rows (field count != header field count).

pub fn with_memory_peak_mb(self, mb: f64) -> Self

Set peak memory usage.

Trait Implementations

impl Clone for ExecutionMetadata

fn clone(&self) -> ExecutionMetadata

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ExecutionMetadata

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for ExecutionMetadata

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for ExecutionMetadata

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

where __S: Serializer,

Serialize this value into the given Serde serializer.