Tables

rerun.blueprint.table

Configure catalog table layouts, columns, and segment previews.

⚠️ The rerun.blueprint.table namespace is unstable and may change significantly.

CardLayout `dataclass`

Display table records as cards.

PARAMETER	DESCRIPTION
`title`	Physical Arrow column name to use for card titles. If unset or not found in the table, uses the first visible string field. TYPE: `str \| None` DEFAULT: `None`
`link`	Physical Arrow column name containing the target to open when a card is activated. If unset or not found in the table, uses the first visible field with a `PreviewCell`. TYPE: `str \| None` DEFAULT: `None`
`fields`	Fields to configure, in display order, with each source listed at most once. Listed fields are visible unless `visible=False`; unlisted fields are hidden. The first visible field with `column_type` set to `rerun.components.Flag` is shown in the card header instead of the labeled-field list. TYPE: `Sequence[Column]` DEFAULT: `()`

Column `dataclass`

Configure a source column's appearance and behavior in one layout.

PARAMETER	DESCRIPTION
`source`	Physical Arrow column name, not a display label. TYPE: `str`
`name`	Display label, overriding `column_display_mode`. If unset, the label is derived from `source`. TYPE: `str \| None` DEFAULT: `None`
`editable`	Allow editing; disabled by default. Requires write permission on a remote table with a column marked by `rerun:is_table_index` metadata. A cell can be edited when its component type has an editor. Boolean columns are shown as `rerun.components.Bool` unless `column_type` is set, e.g. to `rerun.components.Flag`. Editable string and 64-bit float columns without a component type are shown as `rerun.components.Text` and `rerun.components.Scalar`. TYPE: `bool \| None` DEFAULT: `None`
`column_type`	The component the column's values are shown as. Takes a component class like `rerun.components.Url` or `rerun.components.Flag`, a fully qualified component type name like `"rerun.components.Flag"`, or a `ComponentDescriptor` like `rr.TextDocument.descriptor_text()` for the UI of a specific archetype field. The column's Arrow datatype must match the datatype of the component type. `PreviewCell` shows the column as previews instead. If unset, the Viewer uses the component type of the source column, or infers one from the column's datatype. TYPE: `type[ComponentMixin] \| str \| ComponentDescriptor \| PreviewCell \| None` DEFAULT: `None`
`visible`	Override visibility in this layout. Configured columns are visible by default. Set to `False` to hide a column while retaining its configuration and position. TYPE: `bool \| None` DEFAULT: `None`

Filter `dataclass`

A SQL expression that selects the rows of a table.

Build filters for one column with col, which writes the SQL the Viewer shows as an editable filter in the filter bar. Any other SQL expression also filters the table, and the Viewer shows it as SQL text.

A filter has no truth value, so not, and and or raise a TypeError.

PARAMETER	DESCRIPTION
`sql`	A SQL expression that evaluates to a boolean, in the generic SQL dialect of DataFusion. Column names are double-quoted, e.g. `"task" ILIKE '%pick%'`. Functions that combine rows, such as `sum` or `count`, are not allowed. TYPE: `str`

FilterColumn

Bases: _ValueFilters

The column a Filter applies to.

Create it with col.

Comparisons such as col("score") >= 3 and the methods of this class return a Filter. Every filter is one the Viewer shows as an editable filter in its filter bar, when the column has the type the filter is for:

numbers: eq, ne, lt, le, gt and ge, or the operators ==, !=, <, <=, > and >=
booleans: is_true and is_false, and for nullable columns also is_null, is_not_true, is_not_false and is_not_null
strings: str.contains, str.does_not_contain, str.starts_with and str.ends_with
timestamps: before, after, between and the relative ranges such as today, and the negated ranges such as not_between and not_today

The negated filters select the rows where the filter is false or null.

col doesn't know the type of the column, so nothing stops a filter for another type, such as col("score").str.contains("1") on a number column, or col("tags") >= 3 on a list column. The Viewer shows such a filter as SQL text. If the SQL doesn't work on the column, the Viewer shows the error on that filter and doesn't apply it.

For a list column, use the same filters on list.

list `property`

list: ListFilters

The values of a list column, e.g. col("tags").list.str.contains("pick").

str `property`

str: StringFilters

The filters of a string column, e.g. col("task").str.contains("pick").

after

def after(timestamp: datetime) -> Filter

Select rows with a timestamp at or after timestamp.

A datetime without a time zone is in UTC.

before

def before(timestamp: datetime) -> Filter

Select rows with a timestamp before timestamp.

A datetime without a time zone is in UTC.

between

def between(start: datetime, end: datetime) -> Filter

Select rows with a timestamp at or after start and before end.

A datetime without a time zone is in UTC.

eq

def eq(value: Number) -> Filter

Select rows where the value equals value. Same as col(…) == value.

ge

def ge(value: Number) -> Filter

Select rows where the value is at least value. Same as col(…) >= value.

gt

def gt(value: Number) -> Filter

Select rows where the value is greater than value. Same as col(…) > value.

is_false

def is_false() -> Filter

Select rows where the value is false.

is_not_false

def is_not_false() -> Filter

Select rows where the value is true or null.

is_not_null

def is_not_null() -> Filter

Select rows where the value is not null.

For a list column, this selects non-empty lists without nulls.

is_not_true

def is_not_true() -> Filter

Select rows where the value is false or null.

is_null

def is_null() -> Filter

Select rows where the value is null.

For a list column, this selects null rows, empty lists, and lists that contain a null.

is_true

def is_true() -> Filter

Select rows where the value is true.

last_24_hours

def last_24_hours() -> Filter

Select rows with a timestamp in the last 24 hours.

last_week

def last_week() -> Filter

Select rows with a timestamp last week, in the time zone of the Viewer's query session.

le

def le(value: Number) -> Filter

Select rows where the value is at most value. Same as col(…) <= value.

lt

def lt(value: Number) -> Filter

Select rows where the value is less than value. Same as col(…) < value.

ne

def ne(value: Number) -> Filter

Select rows where the value is not value, or is null. Same as col(…) != value.

not_between

def not_between(start: datetime, end: datetime) -> Filter

Select rows the between filter does not select.

not_last_24_hours

def not_last_24_hours() -> Filter

Select rows the last_24_hours filter does not select.

not_last_week

def not_last_week() -> Filter

Select rows the last_week filter does not select.

not_this_week

def not_this_week() -> Filter

Select rows the this_week filter does not select.

not_today

def not_today() -> Filter

Select rows the today filter does not select.

not_yesterday

def not_yesterday() -> Filter

Select rows the yesterday filter does not select.

this_week

def this_week() -> Filter

Select rows with a timestamp this week, in the time zone of the Viewer's query session.

today

def today() -> Filter

Select rows with a timestamp today, in the time zone of the Viewer's query session.

yesterday

def yesterday() -> Filter

Select rows with a timestamp yesterday, in the time zone of the Viewer's query session.

ListFilters

Bases: _ValueFilters

The filters of a list column.

Get them with FilterColumn.list, e.g. col("scores").list >= 3 or col("tags").list.str.contains("pick"). They are the filters of FilterColumn, and a row is selected when the filter holds for any value of its list. The negated filters, such as ne, select the rows where no value matches.

str `property`

str: StringFilters

The filters of a string column, e.g. col("task").str.contains("pick").

after

def after(timestamp: datetime) -> Filter

Select rows with a timestamp at or after timestamp.

A datetime without a time zone is in UTC.

before

def before(timestamp: datetime) -> Filter

Select rows with a timestamp before timestamp.

A datetime without a time zone is in UTC.

between

def between(start: datetime, end: datetime) -> Filter

Select rows with a timestamp at or after start and before end.

A datetime without a time zone is in UTC.

eq

def eq(value: Number) -> Filter

Select rows where the value equals value. Same as col(…) == value.

ge

def ge(value: Number) -> Filter

Select rows where the value is at least value. Same as col(…) >= value.

gt

def gt(value: Number) -> Filter

Select rows where the value is greater than value. Same as col(…) > value.

is_false

def is_false() -> Filter

Select rows where the value is false.

is_not_false

def is_not_false() -> Filter

Select rows where the value is true or null.

is_not_null

def is_not_null() -> Filter

Select rows where the value is not null.

For a list column, this selects non-empty lists without nulls.

is_not_true

def is_not_true() -> Filter

Select rows where the value is false or null.

is_null

def is_null() -> Filter

Select rows where the value is null.

For a list column, this selects null rows, empty lists, and lists that contain a null.

is_true

def is_true() -> Filter

Select rows where the value is true.

last_24_hours

def last_24_hours() -> Filter

Select rows with a timestamp in the last 24 hours.

last_week

def last_week() -> Filter

Select rows with a timestamp last week, in the time zone of the Viewer's query session.

le

def le(value: Number) -> Filter

Select rows where the value is at most value. Same as col(…) <= value.

lt

def lt(value: Number) -> Filter

Select rows where the value is less than value. Same as col(…) < value.

ne

def ne(value: Number) -> Filter

Select rows where the value is not value, or is null. Same as col(…) != value.

not_between

def not_between(start: datetime, end: datetime) -> Filter

Select rows the between filter does not select.

not_last_24_hours

def not_last_24_hours() -> Filter

Select rows the last_24_hours filter does not select.

not_last_week

def not_last_week() -> Filter

Select rows the last_week filter does not select.

not_this_week

def not_this_week() -> Filter

Select rows the this_week filter does not select.

not_today

def not_today() -> Filter

Select rows the today filter does not select.

not_yesterday

def not_yesterday() -> Filter

Select rows the yesterday filter does not select.

this_week

def this_week() -> Filter

Select rows with a timestamp this week, in the time zone of the Viewer's query session.

today

def today() -> Filter

Select rows with a timestamp today, in the time zone of the Viewer's query session.

yesterday

def yesterday() -> Filter

Select rows with a timestamp yesterday, in the time zone of the Viewer's query session.

PreviewCell `dataclass`

Show a recording reference using one or more embedded views.

Preview timing is shared by all preview columns through PreviewsConfig.

PreviewsConfig

Bases: Archetype

Archetype: Shared configuration for every preview cell in a table blueprint.

This archetype is stored at the entity /table alongside archetypes.TableBlueprint.

⚠️ This type is unstable and may change significantly in a way that the data won't be backwards compatible.

init

def __init__(
    *,
    timeline: Utf8Like | None = None,
    play_state: PlayStateLike | None = None,
    playback_speed: Float64Like | None = None,
    start_point: Float64Like | None = None,
) -> None

Create a new instance of the PreviewsConfig archetype.

PARAMETER	DESCRIPTION
`timeline`	The timeline used by every preview cell. If left empty a timeline is automatically picked, preferring custom over built-in. TYPE: `Utf8Like \| None` DEFAULT: `None`
`play_state`	Whether every preview cell plays. A single preview can still be played or paused on its own. Defaults to paused, and following plays like playing. TYPE: `PlayStateLike \| None` DEFAULT: `None`
`playback_speed`	Playback speed of every preview cell. Defaults to 1.0. TYPE: `Float64Like \| None` DEFAULT: `None`
`start_point`	Where every preview cell starts, as a fraction of the recording's time range. 0.0 is the start and 1.0 the end. Defaults to 0.5. TYPE: `Float64Like \| None` DEFAULT: `None`

cleared `classmethod`

def cleared() -> PreviewsConfig

Clear all the fields of a PreviewsConfig.

from_fields `classmethod`

def from_fields(
    *,
    clear_unset: bool = False,
    timeline: Utf8Like | None = None,
    play_state: PlayStateLike | None = None,
    playback_speed: Float64Like | None = None,
    start_point: Float64Like | None = None,
) -> PreviewsConfig

Update only some specific fields of a PreviewsConfig.

PARAMETER	DESCRIPTION
`clear_unset`	If true, all unspecified fields will be explicitly cleared. TYPE: `bool` DEFAULT: `False`
`timeline`	The timeline used by every preview cell. If left empty a timeline is automatically picked, preferring custom over built-in. TYPE: `Utf8Like \| None` DEFAULT: `None`
`play_state`	Whether every preview cell plays. A single preview can still be played or paused on its own. Defaults to paused, and following plays like playing. TYPE: `PlayStateLike \| None` DEFAULT: `None`
`playback_speed`	Playback speed of every preview cell. Defaults to 1.0. TYPE: `Float64Like \| None` DEFAULT: `None`
`start_point`	Where every preview cell starts, as a fraction of the recording's time range. 0.0 is the start and 1.0 the end. Defaults to 0.5. TYPE: `Float64Like \| None` DEFAULT: `None`

SortBy

Encoding: The column and direction a table's rows are sorted by.

⚠️ This type is unstable and may change significantly in a way that the data won't be backwards compatible.

init

def __init__(
    column: Utf8Like,
    *,
    direction: SortDirectionLike | None = None,
) -> None

Create a new instance of the SortBy encoding.

PARAMETER	DESCRIPTION
`column`	The name of the column the rows are sorted by. Same semantics as `components.ColumnName`. TYPE: `Utf8Like`
`direction`	The direction the rows are sorted in. If not set, the viewer sorts ascending. TYPE: `SortDirectionLike \| None` DEFAULT: `None`

StringFilters

The filters of a string column.

Get them with FilterColumn.str, e.g. col("task").str.contains("pick"). The text is matched as plain text, ignoring case, not as a regular expression.

contains

def contains(text: str) -> Filter

Select rows where the value contains text, ignoring case.

does_not_contain

def does_not_contain(text: str) -> Filter

Select rows where the value does not contain text, ignoring case, or is null.

For a list column, this selects rows where no value of the list contains text.

ends_with

def ends_with(text: str) -> Filter

Select rows where the value ends with text, ignoring case.

starts_with

def starts_with(text: str) -> Filter

Select rows where the value starts with text, ignoring case.

TableBlueprint `dataclass`

Configure layouts and segment previews for a catalog table.

This configures a table, not a Viewer viewport Blueprint. Save it as an .rbl file and register its server-accessible URI with TableEntry.register_blueprint(uri) or DatasetEntry.register_blueprint(uri, segment_table=True). See Configure table layouts and segment previews for the full workflow.

PARAMETER	DESCRIPTION
`table_layout`	Row-and-column layout, available by default. TYPE: `TableLayout \| None` DEFAULT: `None`
`card_layout`	Card layout, disabled unless configured. TYPE: `CardLayout \| None` DEFAULT: `None`
`default_layout`	Initial layout: `"table"` or `"cards"`. If unset, uses cards when configured, otherwise the table layout. `"cards"` requires `card_layout`. TYPE: `Literal['table', 'cards'] \| None` DEFAULT: `None`
`column_display_mode`	Column-name formatting in both layouts; defaults to compact formatting. Explicit `Column.name` labels take precedence. TYPE: `ColumnDisplayModeLike \| None` DEFAULT: `None`
`previews_config`	Timeline, play state, and playback speed shared by every `PreviewCell` in both layouts. If no timeline is specified, one is picked automatically, preferring custom over built-in timelines. Previews start paused at 1x speed unless `play_state` or `playback_speed` is set. TYPE: `PreviewsConfig \| None` DEFAULT: `None`
`filters`	Row filters, shared by both layouts. A row is only shown when all filters select it. Build filters for one column with `col`, which the Viewer shows as editable filters. Instead of a `Filter`, you can also provide a string with a raw SQL expression. TYPE: `Sequence[Filter \| str]` DEFAULT: `()`
`sort_by`	Source column the rows are sorted by, with an optional direction. Without a direction the viewer sorts ascending. Use `SortBy` to set the direction. If unset, rows keep the order of the data. TYPE: `str \| SortBy \| None` DEFAULT: `None`

save

def save(
    application_id: str, path: str | Path | None = None
) -> None

Save this table blueprint to an .rbl file for catalog registration.

PARAMETER	DESCRIPTION
`application_id`	Application ID stored in the blueprint recording. TYPE: `str`
`path`	Output file path; defaults to `<application_id>.rbl`. TYPE: `str \| Path \| None` DEFAULT: `None`

TableLayout `dataclass`

Display table records as rows and columns.

PARAMETER	DESCRIPTION
`columns`	Columns to configure, in display order, with each source listed at most once. Listed columns appear first and are visible unless `visible=False`. Unlisted columns follow in default order and retain Viewer visibility defaults. TYPE: `Sequence[Column]` DEFAULT: `()`

col

def col(name: str) -> FilterColumn

Start a filter on the column name.

import rerun.blueprint as rrb
from rerun.blueprint.table import col

rrb.TableBlueprint(
    filters=[
        col("score") >= 3,
        col("success").is_true(),
        col("task").str.contains("pick"),
        col("created").last_week(),
        col("tags").list.does_not_contain("broken"),
    ],
)

For a list column, filter its values with list.

PARAMETER	DESCRIPTION
`name`	Physical Arrow column name. TYPE: `str`

Tables

rerun.blueprint.table

CardLayout dataclass

Column dataclass

Filter dataclass

FilterColumn

list property

str property

after

before

between

eq

ge

gt

is_false

is_not_false

is_not_null

is_not_true

is_null

is_true

last_24_hours

last_week

le

lt

ne

not_between

not_last_24_hours

not_last_week

not_this_week

not_today

not_yesterday

this_week

today

yesterday

ListFilters

str property

after

before

between

eq

ge

gt

is_false

is_not_false

is_not_null

is_not_true

is_null

is_true

last_24_hours

last_week

le

lt

ne

not_between

not_last_24_hours

not_last_week

not_this_week

not_today

not_yesterday

this_week

today

yesterday

PreviewCell dataclass

PreviewsConfig

__init__

cleared classmethod

from_fields classmethod

SortBy

__init__

StringFilters

contains

does_not_contain

ends_with

starts_with

TableBlueprint dataclass

save

TableLayout dataclass

col

CardLayout `dataclass`

Column `dataclass`

Filter `dataclass`

list `property`

str `property`

str `property`

PreviewCell `dataclass`

init

cleared `classmethod`

from_fields `classmethod`

init

TableBlueprint `dataclass`

TableLayout `dataclass`