ncread

Read data from variable in netCDF data source

Syntax

vardata = ncread(source,varname)

vardata = ncread(source,varname,start,count)

vardata = ncread(source,varname,start,count,stride)

Description

vardata = ncread(source,varname) reads all the data from the netCDF variable varname contained in source. The returned variable value vardata is of the MATLAB^® data type that best matches the netCDF data type of varname.

vardata = ncread(source,varname,start,count) reads data beginning at the location specified by start. The count argument specifies the number of elements to read along each dimension.

vardata = ncread(source,varname,start,count,stride) returns data with the interval specified by stride between the indices of each dimension of the variable.

Examples

Read NetCDF Variable

Read and plot a variable named peaks from the netCDF file example.nc.

peaksData = ncread("example.nc","peaks");
whos peaksData

  Name            Size            Bytes  Class    Attributes

  peaksData      50x50             5000  int16              

Plot peaksData and add a title.

surf(peaksData)
title("Peaks Data")

Read Portion of Data from Variable

Read and plot only a subset of the peaks variable data starting from the location [25 17] until the end of each dimension.

start = [25 17];   % Start location along each coordinate
count = [Inf Inf]; % Read until the end of each dimension
peaksData = ncread("example.nc","peaks",start,count);
whos peaksData

  Name            Size            Bytes  Class    Attributes

  peaksData      26x34             1768  int16              

Plot the data.

surf(peaksData)
title("Peaks Data Starting at [25 17]")

Read Data with Specified Spacing Between Variable Indices

Read and plot data, where the data is sampled at a specified spacing between variable indices along each dimension. Read variable data at intervals specified in stride, starting from the location in start. A value of 1 in stride accesses adjacent values in the corresponding dimension, a value of 2 accesses every other value in the corresponding dimension, and so on.

start  = [1 1]; 
count  = [10 15]; 
stride = [2 3]; 
sampledPeaksData = ncread("example.nc","peaks",start,count,stride);
whos sampledPeaksData

  Name                   Size            Bytes  Class    Attributes

  sampledPeaksData      10x15              300  int16              

Plot the data.

surf(sampledPeaksData)
title("Peaks Data Subsampled by [2 3]")

Input Arguments

source — Name of netCDF data source

string scalar | character vector

Name of the netCDF data source, specified as a string scalar or character vector. The source argument can be one of these values:

• The path of a local netCDF source

• The OPeNDAP URL of a remote OPeNDAP netCDF data source

• The HTTP URL of a remote netCDF source, with #mode=bytes appended to the end of the URL to enable byte-range reading

  Note

  Byte-range reading is slower than reading from other sources. For more details about byte-range reading, see the netCDF documentation.

Example: "myNetCDFfile.nc"

Example: "http://host_name/netcdf_filename#mode=bytes"

varname — Variable name

string scalar | character vector

Variable name, specified as a string scalar or character vector containing the name of a variable in the netCDF data source.

If source specifies a file with format netcdf4, you can specify the location of the variable within the group hierarchy by specifying varname as a fully qualified name.

Example: "myVar"

Example: "/myGrp/mySubGrp/myNestedVar"

start — Starting location of data in variable

vector of ones (default) | numeric vector

Starting location of the data in the variable, specified as a numeric vector of positive integers. For an N-dimensional variable, specify start as a vector of length N containing 1-based indices.

If you do not specify start, then the ncread function starts reading the variable from the first index along each dimension.

Example: [2 1 3]

Data Types: double

count — Number of elements

vector of Infs (default) | numeric vector

Number of elements to read, specified as a numeric vector of positive integers or Inf values. For an N-dimensional variable, specify count as a vector of length N, containing the number of elements to read along each dimension. If any element of count is Inf, then ncread reads until the end of the corresponding dimension.

If you do not specify count, then the ncread function reads the variable data until the end of each dimension.

Example: [Inf 10 50]

Data Types: double

stride — Space between variable indices

vector of ones (default) | numeric vector

Space between the variable indices along each dimension, specified as a numeric vector of integers. For an N-dimensional variable, specify stride as a vector of length N. The elements of the stride vector correspond, in order, to the dimensions of the variable. A value of 1 accesses adjacent values of the netCDF variable in the corresponding dimension, a value of 2 accesses every other value in the corresponding dimension, and so on.

If you do not specify stride, then the ncread function reads the data with a default spacing of 1 along each dimension.

Example: [2 10 1]

Data Types: double

Output Arguments

vardata — Variable data

numeric array | text | cell array

Variable data, returned as a numeric array, text, or cell array of the MATLAB data type that best matches the netCDF data type of varname. For more information about how MATLAB determines the best match, see NetCDF to MATLAB Data Type Conversion.

For numeric variables, when at least one of the attributes _FillValue, scale_factor, or add_offset is present, vardata is of type double. Additionally, ncread applies these attribute conventions in a sequence before returning vardata:

1. If the _FillValue attribute exists:

   • If vardata is of type double or single, then ncread replaces vardata values equal to that of the _FillValue attribute with NaN values.

   • If vardata is of any other numeric type, then ncread replaces NaN vardata values as well as vardata values equal to that of the _FillValue attribute with 0 values.

2. If the scale_factor attribute exists, then ncread multiplies the values in vardata by the value of the scale_factor attribute.

3. If the add_offset attribute exists, then ncread adds the value of the add_offset attribute to the values in vardata.

Note

If varname is of type NC_STRING, then it can contain UTF-8-encoded characters; if varname is of type NC_CHAR, then it must contain only ASCII-encoded characters.

Data Types: single | double | int8 | int16 | int32 | int64 | uint8 | uint16 | uint32 | uint64 | cell | char | string

More About

NetCDF to MATLAB Data Type Conversion

The netCDF-related MATLAB functions automatically choose the MATLAB data type that best matches the netCDF data type according to this table.

NetCDF Data Type	MATLAB Data Type
NC_DOUBLE	double
NC_FLOAT	single
NC_INT	int32
NC_SHORT	int16
NC_BYTE	int8
NC_CHAR	char
NC_STRING (*)	string
NC_INT64 (*)	int64
NC_UINT64 (*)	uint64
NC_UINT (*)	uint32
NC_USHORT (*)	uint16
NC_UBYTE (*)	uint8
User-defined NC_VLEN types (*)	cell

(*) These netCDF data types are available only for files with format netcdf4.

Tips

• MATLAB interprets multidimensional data as column-major, but the netCDF C API interprets multidimensional data as row-major. Multidimensional data in the netCDF C API shows dimensions in the reverse of the order shown by MATLAB and consequently appears transposed.

Version History

Introduced in R2011a

R2022a: Byte-range reading of remote datasets

You can use ncread for read-only access to remote datasets using the HTTP byte-range capability, provided that the remote server supports byte-range access.

R2022a: Read variable length array data types (NC_VLEN)

You can read variable length array data types (NC_VLEN) from netCDF-4 files.

R2021b: Read NC_STRING data

You can read NC_STRING data from netCDF-4 files.