Struct rerun::dataframe::external::re_chunk_store::QueryExpression

pub struct QueryExpression {
    pub view_contents: Option<BTreeMap<EntityPath, Option<BTreeSet<ComponentName>>>>,
    pub include_semantically_empty_columns: bool,
    pub include_indicator_columns: bool,
    pub include_tombstone_columns: bool,
    pub filtered_index: Option<Timeline>,
    pub filtered_index_range: Option<ResolvedTimeRange>,
    pub filtered_index_values: Option<BTreeSet<TimeInt>>,
    pub using_index_values: Option<BTreeSet<TimeInt>>,
    pub filtered_is_not_null: Option<ComponentColumnSelector>,
    pub sparse_fill_strategy: SparseFillStrategy,
    pub selection: Option<Vec<ColumnSelector>>,
}

Describes a complete query for Rerun’s dataframe API.

Terminology: view vs. selection vs. filtering vs. sampling

• The view contents specify which subset of the database (i.e., which columns) the query runs on, expressed as a set of EntityPaths and their associated ComponentNames.

• The filters filter out rows of data from the view contents. A filter cannot possibly introduce new rows, it can only remove existing ones from the view contents.

• The samplers sample rows of data from the view contents at user-specified values. Samplers don’t necessarily return existing rows: they might introduce new ones if the sampled value isn’t present in the view contents in the first place.

• The selection applies last and samples columns of data from the filtered/sampled view contents. Selecting a column that isn’t present in the view contents results in an empty column in the final dataframe (null array).

A very rough mental model, in SQL terms:

SELECT <Self::selection> FROM <Self::view_contents> WHERE <Self::filtered_*>

Fields

view_contents: Option<BTreeMap<EntityPath, Option<BTreeSet<ComponentName>>>>

The subset of the database that the query will run on: a set of EntityPaths and their associated ComponentNames.

Defaults to None, which means: everything.

Example (pseudo-code):

view_contents = {
  "world/points": [rr.Position3D, rr.Radius],
  "metrics": [rr.Scalar]
}

include_semantically_empty_columns: bool

Whether the view_contents should ignore semantically empty columns.

A semantically empty column is a column that either contains no data at all, or where all values are either nulls or empty arrays ([]).

view_contents: QueryExpression::view_contents

include_indicator_columns: bool

Whether the view_contents should ignore columns corresponding to indicator components.

Indicator components are marker components, generally automatically inserted by Rerun, that helps keep track of the original context in which a piece of data was logged/sent.

view_contents: QueryExpression::view_contents

include_tombstone_columns: bool

Whether the view_contents should ignore columns corresponding to Clear-related components.

view_contents: QueryExpression::view_contents Clear: re_types_core::archetypes::Clear

filtered_index: Option<Timeline>

The index used to filter out rows from the view contents.

Only rows where at least 1 column contains non-null data at that index will be kept in the final dataset.

If left unspecified, the results will only contain static data.

Examples: Some(Timeline("frame")), None (only static data).

filtered_index_range: Option<ResolvedTimeRange>

The range of index values used to filter out rows from the view contents.

Only rows where at least 1 of the view-contents contains non-null data within that range will be kept in the final dataset.

• This has no effect if filtered_index isn’t set.
• This has no effect if QueryExpression::using_index_values is set.

Example: ResolvedTimeRange(10, 20).

filtered_index_values: Option<BTreeSet<TimeInt>>

The specific index values used to filter out rows from the view contents.

Only rows where at least 1 column contains non-null data at these specific values will be kept in the final dataset.

• This has no effect if filtered_index isn’t set.
• This has no effect if QueryExpression::using_index_values is set.
• Using TimeInt::STATIC as index value has no effect.

Example: [TimeInt(12), TimeInt(14)].

using_index_values: Option<BTreeSet<TimeInt>>

The specific index values used to sample rows from the view contents.

The final dataset will contain one row per sampled index value, regardless of whether data existed for that index value in the view contents. The semantics of the query are consistent with all other settings: the results will be sorted on the filtered_index, and only contain unique index values.

• This has no effect if filtered_index isn’t set.
• If set, this overrides both QueryExpression::filtered_index_range and QueryExpression::filtered_index_values.
• Using TimeInt::STATIC as index value has no effect.

Example: [TimeInt(12), TimeInt(14)].

filtered_is_not_null: Option<ComponentColumnSelector>

The component column used to filter out rows from the view contents.

Only rows where this column contains non-null data be kept in the final dataset.

Example: ComponentColumnSelector("rerun.components.Position3D").

sparse_fill_strategy: SparseFillStrategy

Specifies how null values should be filled in the returned dataframe.

Defaults to SparseFillStrategy::None.

selection: Option<Vec<ColumnSelector>>

The specific columns to sample from the final view contents.

The order of the samples will be respected in the final result.

Defaults to None, which means: everything.

Example: [ColumnSelector(Time("log_time")), ColumnSelector(Component("rerun.components.Position3D"))].

Trait Implementations

impl Clone for QueryExpression

fn clone(&self) -> QueryExpression

Returns a copy of the value.

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

Performs copy-assignment from source.

impl Debug for QueryExpression

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

Formats the value using the given formatter.

impl Default for QueryExpression

fn default() -> QueryExpression

Returns the “default value” for a type.

impl Hash for QueryExpression

fn hash<__H>(&self, state: &mut __H)

where __H: Hasher,

Feeds this value into the given Hasher.

fn hash_slice<H>(data: &[Self], state: &mut H)

where H: Hasher, Self: Sized,

Feeds a slice of this type into the given Hasher.

impl PartialEq for QueryExpression

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

This method tests for self and other values to be equal, and is used by ==.

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

This method tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl TryFrom<Query> for QueryExpression

type Error = TypeConversionError

The type returned in the event of a conversion error.

fn try_from( value: Query ) -> Result<QueryExpression, <QueryExpression as TryFrom<Query>>::Error>

Performs the conversion.

impl Eq for QueryExpression

impl StructuralPartialEq for QueryExpression