How to label an unnamed dimension to access it

[wmcfarlane-smith]

Some data I have doesn't have the time coordinate labelled. It looks like this:
CLOUD CONDENSED ICE COLUMN / (kg/m^2) (-- : 600; latitude: 90; longitude: 144)
Dimension coordinates:
latitude - x -
longitude - - x

The coordinate with 600 points is time but as it has no name I can't access it with cube_1.coords(), only longitude and latitude appear. Is there a way to name time so I can access it? Thanks

[trexfeathers]

Hi @wmcfarlane-smith

The labelling you refer to is the presence of a dimension coordinate on that dimension (more info here: Iris Data Structures). The -- means this is an 'anonymous dimension' - there is no coordinate at all for this dimension. An un-named coordinate for this dimension would instead show as unknown.

You are free to add your own coordinates to an existing Cube, with the following two steps:

1. Create your dimension coordinate. At minimum you will need to include some points for the coordinate, and a name - the name that will appear on the Cube summary.
   For a time coordinate I would also recommend assigning units, to give the numbers some meaning. This can either be a string, like the example below, or a cf-units Unit.
2. Add your new dimension coordinate to your Cube.

from iris.coords import DimCoord
import numpy as np


# Learn more about valid units: https://cf-units.readthedocs.io/en/stable/unit.html
time_coord = DimCoord(np.arange(600), units="days since 1970-01-01")
# Using rename will intelligently assign to the standard_name or long_name property, as appropriate.
time_coord.rename("time")

# 0 refers to which dimension you are adding the coordinate to.
cube_1.add_dim_coord(time_coord, 0)

[wmcfarlane-smith]

Hi Martin, That's really helpful. Thank you!

…

________________________________ From: Martin Yeo ***@***.***> Sent: 14 August 2023 14:18 To: SciTools/iris ***@***.***> Cc: William Smith ***@***.***>; Mention ***@***.***> Subject: Re: [SciTools/iris] How to label an unnamed dimension to access it (Discussion #5422) Hi @wmcfarlane-smith<https://github.com/wmcfarlane-smith> The labelling you refer to is the presence of a dimension coordinate on that dimension (more info here: Iris Data Structures<https://scitools-iris.readthedocs.io/en/stable/userguide/iris_cubes.html>). The -- means this is an 'anonymous dimension' - there is no coordinate at all for this dimension. An un-named coordinate for this dimension would instead show as unknown. You are free to add your own coordinates to an existing Cube, with the following two steps: 1. Create your dimension coordinate. At minimum you will need to include some points for the coordinate, and a name - the name that will appear on the Cube summary. For a time coordinate I would also recommend assigning units, to give the numbers some meaning. This can either be a string, like the example below, or a cf-units Unit<https://cf-units.readthedocs.io/en/stable/unit.html>. 2. Add your new dimension coordinate to your Cube. from iris.coords import DimCoord import numpy as np # Learn more about valid units: https://cf-units.readthedocs.io/en/stable/unit.html time_coord = DimCoord(np.arange(600), units="days since 1970-01-01") # Using rename will intelligently assign to the standard_name or long_name property, as appropriate. time_coord.rename("time") # 0 refers to which dimension you are adding the coordinate to. cube_1.add_dim_coord(time_coord, 0) — Reply to this email directly, view it on GitHub<#5422 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AX4WNMKFCLGEVT4S6R4TJDLXVIQTVANCNFSM6AAAAAA3PTHEPU>. You are receiving this because you were mentioned.Message ID: ***@***.***>