Struct DataQualityPlan 

pub struct DataQualityPlan {

    pub scope: QualityScope,
    pub compute: QualityCompute,
    pub method: SampleMethod,
    pub dataset_rows: usize,
    pub sample_seed: u64,
    pub grain: QualityGrain,
    pub comparison: QualityComparison,
    pub baseline_segment: Option<String>,
    pub temporal_roles: Vec<TemporalRoleAssignment>,
    pub intervals: Option<Vec<(TemporalRole, TemporalRole)>>,
    pub interval_clock: IntervalClock,
    pub latency_threshold_seconds: Option<i64>,
    pub time_formats: Vec<TimeInterpretation>,
    pub expected: Option<ExpectedWindows>,
    pub intent: DeclaredIntent,
}

Fields

scope: QualityScope

compute: QualityCompute

method: SampleMethod

How a dataset-grain sample picks its rows, from the shared analysis sample.

dataset_rows: usize

Rows a dataset-grain sample keeps: the shared analysis sample’s size.

sample_seed: u64

grain: QualityGrain

comparison: QualityComparison

baseline_segment: Option<String>

temporal_roles: Vec<TemporalRoleAssignment>

intervals: Option<Vec<(TemporalRole, TemporalRole)>>

The intervals chosen under Intervals, start role to end role. None until one is chosen: the suggested pairs the assigned roles make.

interval_clock: IntervalClock

Which time puts an interval in a time window.

latency_threshold_seconds: Option<i64>

time_formats: Vec<TimeInterpretation>

Text columns read as time for this study, by grain and roles only.

expected: Option<ExpectedWindows>

The time windows rows are expected in, when stated: what makes a window with no rows a gap. Read from the segments a run counted, so it changes what the report says, never what a run reads.

intent: DeclaredIntent

What the columns must hold, declared: the key and each column’s rules. Read from the rows the run reads; it decides no rows, so it is the report’s.

Implementations

impl DataQualityPlan

pub fn requires_confirmation(&self) -> bool

Only a full scan asks first. Every grain reads the shared sample and cuts it into segments, so no grain reads more than the sample says.

pub fn comparison_label(&self) -> String

pub fn sample(&self) -> Sample

The shared analysis sample this plan carries, as the Sample form shows it.

pub fn adopt_sample(&mut self, sample: &Sample)

Take sample as the rows this plan reads. Metadata-only stays metadata-only; otherwise every row is a full read and anything less a sampled one.

Choosing equal rows per value of a column is choosing to look at that column’s values side by side, and the grain is what does that. Taken only when the choice is new and the grain has not been set, so a grain chosen afterwards stays chosen.

pub fn time_format(&self, column: &str) -> Option<&TimeInterpretation>

How column is read as time, when it is text read through a format.

pub fn time_value(&self, column: &str) -> Expr

A column’s values as time: parsed through its format when it has one, and as stored otherwise.

pub fn reads_as_time(&self, column: &str, schema: &Schema) -> bool

Whether column of schema can be read as time: a date or time type, or text with a format.

pub fn role_column(&self, role: TemporalRole) -> Option<&str>

The column a role is assigned to.

pub fn interval_pairs(&self) -> Vec<(TemporalRole, TemporalRole)>

The intervals this plan measures: the chosen ones, or until one is chosen the suggested pairs; either way only those whose two roles are assigned.

pub fn candidate_pairs(&self) -> Vec<(TemporalRole, TemporalRole)>

Every start and end the assigned roles can make, the suggested pairs first, then the rest in role order: what Intervals in Setup lists.

pub fn toggle_interval(&mut self, pair: (TemporalRole, TemporalRole))

Measure pair, or stop measuring it. The first choice makes the list explicit, starting from what was measured.

pub fn unpaired_roles(&self) -> Vec<TemporalRole>

Assigned roles that no measured interval uses: they measure nothing.

pub fn windows_intervals(&self) -> bool

Whether the clock choice means anything: intervals cut into time windows.

pub fn interval_grain(&self, start: &str, end: &str) -> QualityGrain

The grain an interval from start to end is cut by: the plan’s, except that time windows go by the interval’s own start or end when the clock says.

pub fn zoned(&self, column: &str, schema: &Schema) -> Option<bool>

Whether column holds instants (a zoned type, or text read with an offset) rather than times with no zone; None when it is not read as time.

pub fn same_measurement(&self, other: &Self) -> bool

Whether other measures what this plan measures: they differ, if at all, in the windows they expect, which a report checks against the counts it already holds, or in what its segments are compared with, which is worked out from the segments it holds (DataQualityResults::compare_segments).

pub fn compares_differently(&self, other: &Self) -> bool

Whether other compares segments differently from this plan.

pub fn expected_windows(&self) -> Option<&ExpectedWindows>

The windows this plan expects rows in: only on a time-window grain.

pub fn coarser_grain(&self) -> Option<QualityGrain>

The next coarser grain to offer when segments are thin: a day for an hour, a week for a day, a month for a week, and a larger row chunk. Partitions and files have none.

pub fn set_row_chunks(&mut self)

pub fn compact_summary(&self) -> String

Trait Implementations

impl Clone for DataQualityPlan

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for DataQualityPlan

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

Formats the value using the given formatter.

impl Default for DataQualityPlan

fn default() -> Self

Returns the “default value” for a type.

impl Eq for DataQualityPlan

impl PartialEq for DataQualityPlan

fn eq(&self, other: &Self) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for DataQualityPlan