cfdm.DomainTopology.normalise

DomainTopology.normalise(start_index=0, remove_empty_columns=False, inplace=False)

Normalise the data values.

Normalised data is in a form that is suitable for creating a CF-netCDF UGRID “edge_node_connectivity” or “face_node_connectivity” variable.

Normalisation does not change the logical content of the data. It converts the data so that the set of values comprises all of the integers in the range [0, N-1] (if the start_index parameter is 0), or [1, N] (if start_index is 1), where N is the number of mesh nodes.

New in version (cfdm): 1.11.0.0

Parameters

start_index: int, optional

The start index for the data values in the normalised data. Must be 0 (the default) or 1 for zero- or one-based indices respectively.

remove_empty_columns: bool, optional

If True then remove any array columns that are entirely missing data. By default these are retained.

inplace: bool, optional

If True then do the operation in-place and return None.

Returns

DomainTopology or None

The normalised domain topology construct, or None if the operation was in-place.

**Examples*

Face cells (similarly for edge cells):

>>>

>>> data = cfdm.Data(
...   [[1, 4, 5, 2], [4, 10, 1, -99], [122, 123, 106, 105]],
...   mask_value=-99
... )
>>> d = cfdm.DomainTopology(cell='face', data=data)
>>> print(d.array)
[[1 4 5 2]
 [4 10 1 --]
 [122 123 106 105]]
>>> d0 = d.normalise()
>>> print(c.array)
[[0 2 3 1]
 [2 4 0 --]
 [7 8 6 5]]
>>> (d0.array == d.normalise().array).all()
True
>>> d1 = d.normalise(start_index=1)
>>> print(d1.array)
[[1 3 4 2]
 [3 5 1 --]
 [8 9 7 6]]
>>> (d1.array == d0.array + 1).all()
True

Point cells:

>>>

>>> data = cfdm.Data(
...   [[4, 1, 10, 125], [1, 4, -99, -99], [125, 4, -99, -99]],
...   mask_value=-99
... )
>>> d = cfdm.DomainTopology(cell='point', data=data)
>>> print(d.array)
[[4 1 10 125]
 [1 4 -- --]
 [125 4 -- --]]
>>> print(d.normalise().array)
[[0 1 2]
 [1 0 --]
 [2 0 --]]
>>> print(d.normalise(start_index=1).array)
[[1 2 3]
 [2 1 --]
 [3 1 --]]