ncstruct documentation

The ncstruct function provides an easy syntax to read one or more variables from a netCDF file, either in their entirety or smaller hyperslabs.

Back to Climate Data Tools Contents



Data = ncstruct(file)
Data = ncstruct(file, var1, var2, ...)
Data = ncstruct(file, Scs, ...)


Data = ncstruct(file) reads all variable data from the indicated netCDF file into a structure Data; the field names of Data correspond to the variable names from the file.

Data = ncstruct(file, var1, var2, ...) reads only the specified variable data. The special string 'dimensions' indicates that all dimension variables should be read in; a dimension variable is any variable whose name and size corresponds to one of the file dimensions.

Data = ncstruct(file, Scs, ...) reads data using the dimension hyperslabs in the structure Scs. The field names in Scs should match the dimension names in the file, and each field holds a 1 x 3 array of [start count stride], indicating that data should be read along the specified dimension starting at index start, reading count elements, with inter-element spacing stride.

Example 1: Reading variables from a single file

A small simple file can be read in in its entirety:

A = ncstruct('')
A = 

  struct with fields:

    avagadros_number: 6.0221e+23
         temperature: [50×1 double]
               peaks: [50×50 int16]

For a larger file, you probably don't want or need to read in all the variables. Here, we read in only wind speed variables, along with the dimension variables:

B = ncstruct('', 'u10', 'v10', 'dimensions')
B = 

  struct with fields:

     latitude: [241×1 single]
    longitude: [480×1 single]
         time: [12×1 int32]
          u10: [480×241×12 double]
          v10: [480×241×12 double]

Example 2: Subsetting dimensions

The ncread function allows you to read hyperslabs of variables; however, its syntax requires you to provide start, count, and stride parameters for all dimensions of that variable in the proper order; determining the dimension names and order of dimensions for each variable you want to read can become tedious. The ncstruct function simplfies things by determining which variables include a particular dimension, and in what order, for you.

Returning to the ERA Interim 2017 dataset, we can read data from all variables but only at a single location, and every other time:

Scs = struct('latitude', [1 1 1], 'longitude', [1 1 1], 'time', [1 Inf 2]);
C = ncstruct('', Scs)
C = 

  struct with fields:

    longitude: 0
     latitude: 90
         time: [6×1 int32]
           sp: [1×1×6 double]
          u10: [1×1×6 double]
          v10: [1×1×6 double]
          t2m: [1×1×6 double]
           tp: [1×1×6 double]

Example 3: Reading from multiple files

Multi-file datasets are common in climate model output or other long-running datasets, where data is split into separate files for each day/month/year/etc. to keep file sizes manageable and to allow the dataset to grow as time passes without needing to edit existing files. The following set of files hold one time slice each from the ERA Interim dataset used in the previous examples:

files = {...

We can read data from this set of files using the same syntax as with a single file. Variables that are repeated across the files, such as the dimension variables, will only be read once, while the variables that include the time dimension will be read from each file and concatenated.

D = ncstruct(files, 'u10', 'v10', 'dimensions')
D = 

  struct with fields:

     latitude: [241×1 single]
    longitude: [480×1 single]
         time: [3×1 int32]
          u10: [480×241×3 double]
          v10: [480×241×3 double]

Author Info

This function and supporting documentation was written by Kelly Kearney for the Climate Data Toolbox for Matlab, 2019. It is available as part of this toolbox, and can also be downloaded individually from GitHub.