Logo Iris 1.12

Table Of Contents

Previous topic

1. Introduction

Next topic

3. Saving Iris cubes

This Page

2. Loading Iris cubes

To load a single file into a list of Iris cubes the iris.load() function is used:

import iris
filename = '/path/to/file'
cubes = iris.load(filename)

Iris will attempt to return as few cubes as possible by collecting together multiple fields with a shared standard name into a single multidimensional cube.

The iris.load() function automatically recognises the format of the given files and attempts to produce Iris Cubes from their contents.


Currently there is support for CF NetCDF, GRIB 1 & 2, PP and FieldsFiles file formats with a framework for this to be extended to custom formats.

In order to find out what has been loaded, the result can be printed:

>>> import iris
>>> filename = iris.sample_data_path('uk_hires.pp')
>>> cubes = iris.load(filename)
>>> print(cubes)
0: air_potential_temperature / (K)     (time: 3; model_level_number: 7; grid_latitude: 204; grid_longitude: 187)
1: surface_altitude / (m)              (grid_latitude: 204; grid_longitude: 187)

This shows that there were 2 cubes as a result of loading the file, they were: air_potential_temperature and surface_altitude.

The surface_altitude cube was 2 dimensional with:
  • the two dimensions have extents of 204 and 187 respectively and are represented by the grid_latitude and grid_longitude coordinates.
The air_potential_temperature cubes were 4 dimensional with:
  • the same length grid_latitude and grid_longitude dimensions as surface_altitide
  • a time dimension of length 3
  • a model_level_number dimension of length 7


The result of iris.load() is always a list of cubes. Anything that can be done with a Python list can be done with the resultant list of cubes.


Throughout this user guide you will see the function iris.sample_data_path being used to get the filename for the resources used in the examples. The result of this function is just a string.

Using this function allows us to provide examples which will work across platforms and with data installed in different locations, however in practice you will want to use your own strings:

filename = '/path/to/file'
cubes = iris.load(filename)

To get the air potential temperature cube from the list of cubes returned by iris.load() in the previous example, list indexing can be used:

>>> import iris
>>> filename = iris.sample_data_path('uk_hires.pp')
>>> cubes = iris.load(filename)
>>> # get the first cube (list indexing is 0 based)
>>> air_potential_temperature = cubes[0]
>>> print(air_potential_temperature)
air_potential_temperature / (K)     (time: 3; model_level_number: 7; grid_latitude: 204; grid_longitude: 187)
     Dimension coordinates:
          time                           x                      -                 -                    -
          model_level_number             -                      x                 -                    -
          grid_latitude                  -                      -                 x                    -
          grid_longitude                 -                      -                 -                    x
     Auxiliary coordinates:
          forecast_period                x                      -                 -                    -
          level_height                   -                      x                 -                    -
          sigma                          -                      x                 -                    -
          surface_altitude               -                      -                 x                    x
     Derived coordinates:
          altitude                       -                      x                 x                    x
     Scalar coordinates:
          forecast_reference_time: 2009-11-19 04:00:00
          STASH: m01s00i004
          source: Data from Met Office Unified Model
          um_version: 7.3

Notice that the result of printing a cube is a little more verbose than it was when printing a list of cubes. In addition to the very short summary which is provided when printing a list of cubes, information is provided on the coordinates which constitute the cube in question. This was the output discussed at the end of the Introduction section.


Dimensioned coordinates will have a dimension marker x in the appropriate column for each cube data dimension that they describe.

2.1. Loading multiple files

To load more than one file into a list of cubes, a list of filenames can be provided to iris.load():

filenames = [iris.sample_data_path('uk_hires.pp'),
cubes = iris.load(filenames)

It is also possible to load one or more files with wildcard substitution using the expansion rules defined fnmatch.

For example, to match zero or more characters in the filename, star wildcards can be used:

filename = iris.sample_data_path('GloSea4', '*.pp')
cubes = iris.load(filename)

2.2. Constrained loading

Given a large dataset, it is possible to restrict or constrain the load to match specific Iris cube metadata. Constrained loading provides the ability to generate a cube from a specific subset of data that is of particular interest.

As we have seen, loading the following file creates several Cubes:

filename = iris.sample_data_path('uk_hires.pp')
cubes = iris.load(filename)

Specifying a name as a constraint argument to iris.load() will mean only cubes with a matching name will be returned:

filename = iris.sample_data_path('uk_hires.pp')
cubes = iris.load(filename, 'specific_humidity')

To constrain the load to multiple distinct constraints, a list of constraints can be provided. This is equivalent to running load once for each constraint but is likely to be more efficient:

filename = iris.sample_data_path('uk_hires.pp')
cubes = iris.load(filename, ['air_potential_temperature', 'specific_humidity'])

The iris.Constraint class can be used to restrict coordinate values on load. For example, to constrain the load to match a specific model_level_number:

filename = iris.sample_data_path('uk_hires.pp')
level_10 = iris.Constraint(model_level_number=10)
cubes = iris.load(filename, level_10)

Constraints can be combined using & to represent a more restrictive constraint to load:

filename = iris.sample_data_path('uk_hires.pp')
forecast_6 = iris.Constraint(forecast_period=6)
level_10 = iris.Constraint(model_level_number=10)
cubes = iris.load(filename, forecast_6 & level_10)

As well as being able to combine constraints using &, the iris.Constraint class can accept multiple arguments, and a list of values can be given to constrain a coordinate to one of a collection of values:

filename = iris.sample_data_path('uk_hires.pp')
level_10_or_16_fp_6 = iris.Constraint(model_level_number=[10, 16], forecast_period=6)
cubes = iris.load(filename, level_10_or_16_fp_6)

A common requirement is to limit the value of a coordinate to a specific range, this can be achieved by passing the constraint a function:

def bottom_16_levels(cell):
   # return True or False as to whether the cell in question should be kept
   return cell <= 16

filename = iris.sample_data_path('uk_hires.pp')
level_lt_16 = iris.Constraint(model_level_number=bottom_16_levels)
cubes = iris.load(filename, level_lt_16)


As with many of the examples later in this documentation, the simple function above can be conveniently written as a lambda function on a single line:

bottom_16_levels = lambda cell: cell <= 16

Note also the warning on equality constraints with floating point coordinates.

Cube attributes can also be part of the constraint criteria. Supposing a cube attribute of STASH existed, as is the case when loading PP files, then specific STASH codes can be filtered:

filename = iris.sample_data_path('uk_hires.pp')
level_10_with_stash = iris.AttributeConstraint(STASH='m01s00i004') & iris.Constraint(model_level_number=10)
cubes = iris.load(filename, level_10_with_stash)

See also

For advanced usage there are further examples in the iris.Constraint reference documentation.

2.2.1. Constraining on Time

Iris follows NetCDF-CF rules in representing time coordinate values as normalised, purely numeric, values which are normalised by the calendar specified in the coordinate’s units (e.g. “days since 1970-01-01”). However, when constraining by time we usually want to test calendar-related aspects such as hours of the day or months of the year, so Iris provides special features to facilitate this:

Firstly, Iris can be configured so that when it evaluates Constraint expressions, it will convert time-coordinate values (points and bounds) from numbers into datetime-like objects for ease of calendar-based testing. This feature is not backwards compatible so for now it must be explicitly enabled by setting the “cell_datetime_objects” option in iris.Future:

>>> filename = iris.sample_data_path('uk_hires.pp')
>>> cube_all = iris.load_cube(filename, 'air_potential_temperature')
>>> print('All times :\n' + str(cube_all.coord('time')))
All times :
DimCoord([2009-11-19 10:00:00, 2009-11-19 11:00:00, 2009-11-19 12:00:00], standard_name='time', calendar='gregorian')
>>> # Define a function which accepts a datetime as its argument (this is simplified in later examples).
>>> hour_11 = iris.Constraint(time=lambda cell: cell.point.hour == 11)
>>> with iris.FUTURE.context(cell_datetime_objects=True):
...     cube_11 = cube_all.extract(hour_11)
>>> print('Selected times :\n' + str(cube_11.coord('time')))
Selected times :
DimCoord([2009-11-19 11:00:00], standard_name='time', calendar='gregorian')

Secondly, the iris.time module provides flexible time comparison facilities. An iris.time.PartialDateTime object can be compared to objects such as datetime.datetime instances, and this comparison will then test only those ‘aspects’ which the PartialDateTime instance defines:

>>> import datetime
>>> from iris.time import PartialDateTime
>>> dt = datetime.datetime(2011, 3, 7)
>>> print(dt > PartialDateTime(year=2010, month=6))
>>> print(dt > PartialDateTime(month=6))

These two facilities can be combined to provide straightforward calendar-based time selections when loading or extracting data.

The previous constraint example can now be written as:

>>> the_11th_hour = iris.Constraint(time=iris.time.PartialDateTime(hour=11))
>>> with iris.FUTURE.context(cell_datetime_objects=True):
...     print(iris.load_cube(
...         iris.sample_data_path('uk_hires.pp'),
...         'air_potential_temperature' & the_11th_hour).coord('time'))
DimCoord([2009-11-19 11:00:00], standard_name='time', calendar='gregorian')

A more complex example might be when there exists a time sequence representing the first day of every week for many years:

>>> print(long_ts.coord('time'))
DimCoord([2007-04-09 00:00:00, 2007-04-16 00:00:00, 2007-04-23 00:00:00,
          2010-02-01 00:00:00, 2010-02-08 00:00:00, 2010-02-15 00:00:00],
         standard_name='time', calendar='gregorian')

We can select points within a certain part of the year, in this case between the 15th of July through to the 25th of August, by combining the datetime cell functionality with PartialDateTime:

>>> st_swithuns_daterange = iris.Constraint(
...     time=lambda cell: PartialDateTime(month=7, day=15) < cell < PartialDateTime(month=8, day=25))
>>> with iris.FUTURE.context(cell_datetime_objects=True):
...   within_st_swithuns = long_ts.extract(st_swithuns_daterange)
>>> print(within_st_swithuns.coord('time'))
DimCoord([2007-07-16 00:00:00, 2007-07-23 00:00:00, 2007-07-30 00:00:00,
       2007-08-06 00:00:00, 2007-08-13 00:00:00, 2007-08-20 00:00:00,
       2008-07-21 00:00:00, 2008-07-28 00:00:00, 2008-08-04 00:00:00,
       2008-08-11 00:00:00, 2008-08-18 00:00:00, 2009-07-20 00:00:00,
       2009-07-27 00:00:00, 2009-08-03 00:00:00, 2009-08-10 00:00:00,
       2009-08-17 00:00:00, 2009-08-24 00:00:00], standard_name='time', calendar='gregorian')

Notice how the dates printed are between the range specified in the st_swithuns_daterange and that they span multiple years.

2.3. Strict loading

The iris.load_cube() and iris.load_cubes() functions are similar to iris.load() except they can only return one cube per constraint. The iris.load_cube() function accepts a single constraint and returns a single cube. The iris.load_cubes() function accepts any number of constraints and returns a list of cubes (as an iris.cube.CubeList). Providing no constraints to iris.load_cube() or iris.load_cubes() is equivalent to requesting exactly one cube of any type.

A single cube is loaded in the following example:

>>> filename = iris.sample_data_path('air_temp.pp')
>>> cube = iris.load_cube(filename)
>>> print(cube)
air_temperature / (K)                 (latitude: 73; longitude: 96)
     Dimension coordinates:
          latitude                           x              -
          longitude                          -              x
     Cell methods:
          mean: time

However, when attempting to load data which would result in anything other than one cube, an exception is raised:

>>> filename = iris.sample_data_path('uk_hires.pp')
>>> cube = iris.load_cube(filename)
Traceback (most recent call last):
iris.exceptions.ConstraintMismatchError: Expected exactly one cube, found 2.


All the load functions share many of the same features, hence multiple files could be loaded with wildcard filenames or by providing a list of filenames.

The strict nature of iris.load_cube() and iris.load_cubes() means that, when combined with constrained loading, it is possible to ensure that precisely what was asked for on load is given - otherwise an exception is raised. This fact can be utilised to make code only run successfully if the data provided has the expected criteria.

For example, suppose that code needed air_potential_temperature in order to run:

import iris
filename = iris.sample_data_path('uk_hires.pp')
air_pot_temp = iris.load_cube(filename, 'air_potential_temperature')

Should the file not produce exactly one cube with a standard name of ‘air_potential_temperature’, an exception will be raised.

Similarly, supposing a routine needed both ‘surface_altitude’ and ‘air_potential_temperature’ to be able to run:

import iris
filename = iris.sample_data_path('uk_hires.pp')
altitude_cube, pot_temp_cube = iris.load_cubes(filename, ['surface_altitude', 'air_potential_temperature'])

The result of iris.load_cubes() in this case will be a list of 2 cubes ordered by the constraints provided. Multiple assignment has been used to put these two cubes into separate variables.


In Python, lists of a pre-known length and order can be exploited using multiple assignment:

>>> number_one, number_two = [1, 2]
>>> print(number_one)
>>> print(number_two)