# xarray.set_options
### *class* xarray.set_options(\*\*kwargs)[[source]](https://github.com/pydata/xarray/blob/main/xarray/core/options.py#L185-L446)
Set options for xarray in a controlled context.
* **Parameters:**
* **arithmetic_broadcast** ([`bool`](https://docs.python.org/3/builtins/functions.html#bool), *default*: [`True`](https://docs.python.org/3/builtins/constants.html#True)) – Whether to perform automatic broadcasting in binary operations.
* **arithmetic_compat** (`{"identical", "equals", "broadcast_equals", "no_conflicts", "override", "minimal"}`, *default*: `"minimal"`) – How to compare non-index coordinates of the same name for potential
conflicts when performing binary operations. (For the alignment of index
coordinates in binary operations, see arithmetic_join.)
Allowed values are:
- `"identical"`: all values, dimensions and attributes of the coordinates
must be the same.
- `"equals"`: all values and dimensions of the coordinates must be the
same.
- `"broadcast_equals"`: all values of the coordinates must be equal after
broadcasting to ensure common dimensions.
- `"no_conflicts"`: only values which are not null in both coordinates
must be equal. The returned coordinate then contains the combination
of all non-null values.
- `"override"`: skip comparing and take the coordinates from the first
operand.
- `"minimal"`: drop conflicting coordinates.
* **arithmetic_join** (`{"inner", "outer", "left", "right", "exact"}`, *default*: `"inner"`) – DataArray/Dataset index alignment in binary operations.
Allowed values are:
- `"outer"`: use the union of object indexes
- `"inner"`: use the intersection of object indexes
- `"left"`: use indexes from the first object with each dimension
- `"right"`: use indexes from the last object with each dimension
- `"exact"`: instead of aligning, raise ValueError when indexes to be
aligned are not equal
* **chunk_manager** ([`str`](https://docs.python.org/3/builtins/stdtypes.html#str), *default*: `"dask"`) – Chunk manager to use for chunked array computations when multiple
options are installed.
* **cmap_divergent** ([`str`](https://docs.python.org/3/builtins/stdtypes.html#str) or [`matplotlib.colors.Colormap`](https://matplotlib.org/stable/api/_as_gen/matplotlib.colors.Colormap.html#matplotlib.colors.Colormap), *default*: `"RdBu_r"`) – Colormap to use for divergent data plots. If string, must be a
matplotlib built-in colormap. Can also be a Colormap object
(e.g. `mpl.colormaps["magma"]`).
* **cmap_sequential** ([`str`](https://docs.python.org/3/builtins/stdtypes.html#str) or [`matplotlib.colors.Colormap`](https://matplotlib.org/stable/api/_as_gen/matplotlib.colors.Colormap.html#matplotlib.colors.Colormap), *default*: `"viridis"`) – Colormap to use for non-divergent data plots. If string, must be a
matplotlib built-in colormap. Can also be a Colormap object
(e.g. `mpl.colormaps["magma"]`).
* **display_default_indexes** (`{"default", True, False}`, *default*: [`False`](https://docs.python.org/3/builtins/constants.html#False)) – Whether to show default indexes, i.e. `PandasIndex` and
`PandasMultiIndex`, in the indexes section of the repr of
`DataArray`, `Dataset` and `DataTree` objects. `"default"` hides
them, like `False`.
* **display_expand_attrs** (`{"default", True, False}`, *default*: `"default"`) – Whether to expand the attributes section for display of
`DataArray` or `Dataset` objects.
Allowed values are:
* `True` : always expand attrs
* `False` : always collapse attrs
* `"default"` : expand unless over a pre-defined limit
* **display_expand_coords** (`{"default", True, False}`, *default*: `"default"`) – Whether to expand the coordinates section for display of
`DataArray` or `Dataset` objects.
Allowed values are:
* `True` : always expand coordinates
* `False` : always collapse coordinates
* `"default"` : expand unless over a pre-defined limit
* **display_expand_data** (`{"default", True, False}`, *default*: `"default"`) – Whether to expand the data section for display of `DataArray`
objects.
Allowed values are:
* `True` : always expand data
* `False` : always collapse data
* `"default"` : expand unless over a pre-defined limit
* **display_expand_data_vars** (`{"default", True, False}`, *default*: `"default"`) – Whether to expand the data variables section for display of
`Dataset` objects.
Allowed values are:
* `True` : always expand data variables
* `False` : always collapse data variables
* `"default"` : expand unless over a pre-defined limit
* **display_expand_groups** (`{"default", True, False}`, *default*: `"default"`) – Has no effect since the HTML repr of `DataTree` was reworked in
v2026.01.0. Use `display_max_children` and
`display_max_html_elements` to limit the displayed groups instead.
* **display_expand_indexes** (`{"default", True, False}`, *default*: `"default"`) – Whether to expand the indexes section for display of
`DataArray` or `Dataset` objects.
Allowed values are:
* `True` : always expand indexes
* `False` : always collapse indexes
* `"default"` : expand unless over a pre-defined limit
(always collapse for HTML style)
* **display_max_children** ([`int`](https://docs.python.org/3/builtins/functions.html#int), *default*: `12`) – Maximum number of children to display for each node in a DataTree.
* **display_max_html_elements** ([`int`](https://docs.python.org/3/builtins/functions.html#int), *default*: `300`) – Maximum number of HTML elements to include in DataTree HTML displays.
Additional items are truncated.
* **display_max_rows** ([`int`](https://docs.python.org/3/builtins/functions.html#int), *default*: `12`) – Maximum display rows.
* **display_max_items** ([`int`](https://docs.python.org/3/builtins/functions.html#int), *default*: `20`) – Maximum number of items to display for a DataTree before collapsing
child nodes, across all levels.
* **display_values_threshold** ([`int`](https://docs.python.org/3/builtins/functions.html#int), *default*: `200`) – Total number of array elements which trigger summarization rather
than full repr for variable data views (NumPy arrays).
* **display_style** (`{"text", "html"}`, *default*: `"html"`) – Display style to use in Jupyter for xarray objects.
* **display_width** ([`int`](https://docs.python.org/3/builtins/functions.html#int), *default*: `80`) – Maximum display width for `repr` on xarray objects.
* **facetgrid_figsize** (`{"computed", "rcparams"}` or [`tuple`](https://docs.python.org/3/builtins/stdtypes.html#tuple) of [`float`](https://docs.python.org/3/builtins/functions.html#float), *default*: `"computed"`) – How [`FacetGrid`](xarray.plot.FacetGrid.html.md#xarray.plot.FacetGrid) determines figure size when
`figsize` is not explicitly passed:
* `"computed"` : figure size is derived from `size` and `aspect`
parameters (current default behavior).
* `"rcparams"` : use `matplotlib.rcParams['figure.figsize']` as the
total figure size.
* `(width, height)` : use a fixed figure size (in inches).
* **file_cache_maxsize** ([`int`](https://docs.python.org/3/builtins/functions.html#int), *default*: `128`) – Maximum number of open files to hold in xarray’s global
least-recently-used cache. This should be smaller than your system’s
per-process file descriptor limit (e.g., `ulimit -n` on Linux).
* **keep_attrs** (`{"default", True, False}`, *default*: `"default"`) – Whether to keep attributes on xarray Datasets/DataArrays after
operations.
Allowed values are:
* `True` : always keep attrs
* `False` : always discard attrs
* `"default"` : keep attrs only in unambiguous circumstances
* **netcdf_engine_order** ([sequence](https://docs.python.org/3/glossary.html#term-sequence) of `{"netcdf4", "h5netcdf", "scipy"}`) – Preference order of backend engines to use when reading or writing
netCDF files with `open_dataset()` and `to_netcdf()` if `engine`
is not explicitly specified. May be any permutation or subset of the
allowed values.
Default is `("netcdf4", "h5netcdf", "scipy")`.
* **use_bottleneck** ([`bool`](https://docs.python.org/3/builtins/functions.html#bool), *default*: [`False`](https://docs.python.org/3/builtins/constants.html#False)) – Whether to use `bottleneck` to accelerate 1D reductions and
1D rolling reduction operations.
* **use_flox** ([`bool`](https://docs.python.org/3/builtins/functions.html#bool), *default*: [`True`](https://docs.python.org/3/builtins/constants.html#True)) – Whether to use `numpy_groupies` and `flox` to accelerate
groupby and resampling reductions.
* **use_new_combine_kwarg_defaults** ([`bool`](https://docs.python.org/3/builtins/functions.html#bool), *default*: [`False`](https://docs.python.org/3/builtins/constants.html#False)) – Whether to use new kwarg default values for combine functions:
[`concat()`](xarray.concat.html.md#xarray.concat), [`merge()`](xarray.merge.html.md#xarray.merge),
[`open_mfdataset()`](xarray.open_mfdataset.html.md#xarray.open_mfdataset). New values are:
* `data_vars`: None
* `coords`: “minimal”
* `compat`: “override”
* `join`: “exact”
* **use_numbagg** ([`bool`](https://docs.python.org/3/builtins/functions.html#bool), *default*: [`True`](https://docs.python.org/3/builtins/constants.html#True)) – Whether to use `numbagg` to accelerate reductions.
Takes precedence over `use_bottleneck` when both are True.
* **use_opt_einsum** ([`bool`](https://docs.python.org/3/builtins/functions.html#bool), *default*: [`True`](https://docs.python.org/3/builtins/constants.html#True)) – Whether to use `opt_einsum` to accelerate dot products.
* **warn_for_unclosed_files** ([`bool`](https://docs.python.org/3/builtins/functions.html#bool), *default*: [`False`](https://docs.python.org/3/builtins/constants.html#False)) – Whether or not to issue a warning when unclosed files are deallocated.
This is mostly useful for debugging.
### Examples
It is possible to use `set_options` either as a context manager:
```pycon
>>> ds = xr.Dataset({"x": np.arange(1000)})
>>> with xr.set_options(display_width=40):
... print(ds)
...
Size: 8kB
Dimensions: (x: 1000)
Coordinates:
* x (x) int64 8kB 0 1 ... 999
Data variables:
*empty*
```
Or to set global options:
```pycon
>>> xr.set_options(display_width=80)
```