pyvista.DataSetFilters.extract_values

DataSetFilters.extract_values(

values: float | VectorLike[float] | MatrixLike[float] | dict[str, float] | dict[float, str] | None = None,

*,

ranges: VectorLike[float] | MatrixLike[float] | dict[str, VectorLike[float]] | dict[tuple[float, float], str] | None = None,

scalars: str | None = None,

preference: Literal['point', 'cell'] = 'point',

component_mode: Literal['any', 'all', 'multi'] | int = 'all',

invert: bool = False,

adjacent_cells: bool = True,

include_cells: bool | None = None,

split: bool = False,

pass_point_ids: bool = True,

pass_cell_ids: bool = True,

progress_bar: bool = False,

)

Return a subset of the mesh based on the value(s) of point or cell data.

Points and cells may be extracted with a single value, multiple values, a range of values, or any mix of values and ranges. This enables threshold-like filtering of data in a discontinuous manner to extract a single label or groups of labels from categorical data, or to extract multiple regions from continuous data. Extracted values may optionally be split into separate meshes.

This filter operates on point data and cell data distinctly:

Point data

All cells with at least one point with the specified value(s) are returned. Optionally, set adjacent_cells to False to only extract points from cells where all points in the cell strictly have the specified value(s). In these cases, a point is only included in the output if that point is part of an extracted cell.

Alternatively, set include_cells to False to exclude cells from the operation completely and extract all points with a specified value.

Cell Data

Only the cells (and their points) with the specified values(s) are included in the output.

Internally, extract_points() is called to extract points for point data, and extract_cells() is called to extract cells for cell data.

By default, two arrays are included with the output: 'vtkOriginalPointIds' and 'vtkOriginalCellIds'. These arrays can be used to link the filtered points or cells directly to the input.

Added in version 0.44.

Parameters:

valuesfloat | ArrayLike[float] | dict, optional

Value(s) to extract. Can be a number, an iterable of numbers, or a dictionary with numeric entries. For dict inputs, either its keys or values may be numeric, and the other field must be strings. The numeric field is used as the input for this parameter, and if split is True, the string field is used to set the block names of the returned MultiBlock.

Note

When extracting multi-component values with component_mode=multi, each value is specified as a multi-component scalar. In this case, values can be a single vector or an array of row vectors.

rangesArrayLike[float] | dict, optional

Range(s) of values to extract. Can be a single range (i.e. a sequence of two numbers in the form [lower, upper]), a sequence of ranges, or a dictionary with range entries. Any combination of values and ranges may be specified together. The endpoints of the ranges are included in the extraction. Ranges cannot be set when component_mode=multi.

For dict inputs, either its keys or values may be numeric, and the other field must be strings. The numeric field is used as the input for this parameter, and if split is True, the string field is used to set the block names of the returned MultiBlock.

Note

Use +/- infinity to specify an unlimited bound, e.g.:

• [0, float('inf')] to extract values greater than or equal to zero.

• [float('-inf'), 0] to extract values less than or equal to zero.

scalarsstr, optional

Name of scalars to extract with. Defaults to currently active scalars.

preferencestr, default: ‘point’

When scalars is specified, this is the preferred array type to search for in the dataset. Must be either 'point' or 'cell'.

component_modeint | ‘any’ | ‘all’ | ‘multi’, default: ‘all’

Specify the component(s) to use when scalars is a multi-component array. Has no effect when the scalars have a single component. Must be one of:

• number: specify the component number as a 0-indexed integer. The selected component must have the specified value(s).

• 'any': any single component can have the specified value(s).

• 'all': all individual components must have the specified values(s).

• 'multi': the entire multi-component item must have the specified value.

invertbool, default: False

Invert the extraction values. If True extract the points (with cells) which do not have the specified values.

adjacent_cellsbool, default: True

If True, include cells (and their points) that contain at least one of the extracted points. If False, only include cells that contain exclusively points from the extracted points list. Has no effect if include_cells is False. Has no effect when extracting values from cell data.

include_cellsbool, default: None

Specify if cells shall be used for extraction or not. If False, points with the specified values are extracted regardless of their cell connectivity, and all cells at the output will be vertex cells (one for each point.) Has no effect when extracting values from cell data.

By default, this value is True if the input has at least one cell and False otherwise.

splitbool, default: False

If True, each value in values and each range in range is extracted independently and returned as a MultiBlock. The number of blocks returned equals the number of input values and ranges. The blocks may be named if a dictionary is used as input. See values and ranges for details.

Note

Output blocks may contain empty meshes if no values meet the extraction criteria. This can impact plotting since empty meshes cannot be plotted by default. Use pyvista.MultiBlock.clean() on the output to remove empty meshes, or set pv.global_theme.allow_empty_mesh = True to enable plotting empty meshes.

pass_point_idsbool, default: True

Add a point array 'vtkOriginalPointIds' that identifies the original points the extracted points correspond to.

pass_cell_idsbool, default: True

Add a cell array 'vtkOriginalCellIds' that identifies the original cells the extracted cells correspond to.

progress_barbool, default: False

Display a progress bar to indicate progress.

Returns:

pyvista.UnstructuredGrid or pyvista.MultiBlock

An extracted mesh or a composite of extracted meshes, depending on split.

See also

split_values

Wrapper around this filter to split values and return a MultiBlock.

select_values()

Similar filter specialized for ImageData.

extract_points

Extract a subset of a mesh’s points.

extract_cells

Extract a subset of a mesh’s cells.

threshold

Similar filter for thresholding a mesh by value.

partition

Split a mesh into a number of sub-parts.

Examples

Extract a single value from a grid’s point data.

>>> import numpy as np
>>> import pyvista as pv
>>> from pyvista import examples
>>> mesh = examples.load_uniform()
>>> extracted = mesh.extract_values(0)

Plot extracted values. Since adjacent cells are included by default, points with values other than 0 are included in the output.

>>> extracted.get_data_range()
(np.float64(0.0), np.float64(81.0))
>>> extracted.plot()

Static Scene

Interactive Scene

Set include_cells=False to only extract points. The output scalars now strictly contain zeros.

>>> extracted = mesh.extract_values(0, include_cells=False)
>>> extracted.get_data_range()
(np.float64(0.0), np.float64(0.0))
>>> extracted.plot(render_points_as_spheres=True, point_size=100)

Static Scene

Interactive Scene

Use ranges to extract values from a grid’s point data in range.

Here, we use +/- infinity to extract all values of 100 or less.

>>> extracted = mesh.extract_values(ranges=[-np.inf, 100])
>>> extracted.plot()

Static Scene

Interactive Scene

Extract every third cell value from cell data.

>>> mesh = examples.load_hexbeam()
>>> lower, upper = mesh.get_data_range()
>>> step = 3
>>> extracted = mesh.extract_values(
...     range(lower, upper, step)  # values 0, 3, 6, ...
... )

Plot result and show an outline of the input for context.

>>> pl = pv.Plotter()
>>> _ = pl.add_mesh(extracted)
>>> _ = pl.add_mesh(mesh.extract_all_edges())
>>> pl.show()

Static Scene

Interactive Scene

Any combination of values and ranges may be specified.

E.g. extract a single value and two ranges, and split the result into separate blocks of a MultiBlock.

>>> extracted = mesh.extract_values(
...     values=18, ranges=[[0, 8], [29, 40]], split=True
... )
>>> extracted
MultiBlock (...)
  N Blocks:   3
  X Bounds:   0.000e+00, 1.000e+00
  Y Bounds:   0.000e+00, 1.000e+00
  Z Bounds:   0.000e+00, 5.000e+00
>>> extracted.plot(multi_colors=True)

Static Scene

Interactive Scene

Extract values from multi-component scalars.

First, create a point cloud with a 3-component RGB color array.

>>> rng = np.random.default_rng(seed=1)
>>> points = rng.random((30, 3))
>>> colors = rng.random((30, 3))
>>> point_cloud = pv.PointSet(points)
>>> point_cloud['colors'] = colors
>>> plot_kwargs = dict(render_points_as_spheres=True, point_size=50, rgb=True)
>>> point_cloud.plot(**plot_kwargs)

Static Scene

Interactive Scene

Extract values from a single component.

E.g. extract points with a strong red component (i.e. > 0.8).

>>> extracted = point_cloud.extract_values(ranges=[0.8, 1.0], component_mode=0)
>>> extracted.plot(**plot_kwargs)

Static Scene

Interactive Scene

Extract values from all components.

E.g. extract points where all RGB components are dark (i.e. < 0.5).

>>> extracted = point_cloud.extract_values(
...     ranges=[0.0, 0.5], component_mode='all'
... )
>>> extracted.plot(**plot_kwargs)

Static Scene

Interactive Scene

Extract specific multi-component values.

E.g. round the scalars to create binary RGB components, and extract only green and blue components.

>>> point_cloud['colors'] = np.round(point_cloud['colors'])
>>> green = [0, 1, 0]
>>> blue = [0, 0, 1]
>>>
>>> extracted = point_cloud.extract_values(
...     values=[blue, green],
...     component_mode='multi',
... )
>>> extracted.plot(**plot_kwargs)

Static Scene

Interactive Scene

Use the original IDs returned by the extraction to modify the original point cloud.

For example, change the color of the blue and green points to yellow.

>>> point_ids = extracted['vtkOriginalPointIds']
>>> yellow = [1, 1, 0]
>>> point_cloud['colors'][point_ids] = yellow
>>> point_cloud.plot(**plot_kwargs)

Static Scene

Interactive Scene