4. A complete description of netCDF NCA attributes describing a master data array

netCDF properties	nca_array attributes	Partitions properties	subarray properties
cf_role, nca_dimensions, nca_array	directions, pmdimensions, pmdimensions, pmshape, base, Partitions,	index, location, pdimensions, pdirections, format, units, calendar, part, subarray	pshape, file, pdtype, ncvar, file_offset, scale_factor, add_offset, lbpack

4.1. cf_role

cf_role

This standard CF property must be used to mark the netCDF variable as either an NCA variable or a private NCA variable.

It takes values of "nca_variable" or "nca_private" for an NCA variable or a NCA private variable respectively.

Examples
"nca_variable"
"nca_private"

4.2. nca_dimensions

nca_dimensions

A netCDF property whose value is a string containing the ordered, space delimited set of the master array’s dimension names. The specified dimensions are all defined as netCDF dimensions in the NCA file.

If missing, an empty sting or a string containing only spaces then the master array is assumed to be scalar.

Examples
"time lat lon"
""
"  "

The master array’s dimensions are stored outside of the nca_array attribute in order to provide useful information about the master array without having to decode the nca_array string (which may not be possible in all APIs).

4.3. nca_array

nca_array

A netCDF property whose value is a JSON encoded string containing parameters required for constructing the aggregated data array.

The dimensions, shape, dtype, units and calendar parameters are specified elsewhere (by attributes or properties of the netCDF NCA variable or inferrable from the netCDF dimensions in NCA file) and so are not required.

Examples
"{'directions': {'lat': false, 'lon': true }, 'Partitions':[{'index': [0], 'data': {'file': 'test1.nc', 'shape': [64, 128], 'ncvar': 'tas'}, 'location': [[0, 64], [0, 128]], 'format': 'netCDF'}]}"

The JSON decoded parameters are described here:

directions

An associative array mapping each dimension of the master array to a direction. Each direction is described as true (increasing) or false (decreasing).

If the master array is a scalar then a boolean rather than an associative array is given.

Examples
{'time', true, 'lat': false, 'lon', true}
true
false

Note

A scalar master array may have an implied direction, e.g. if there are bounds associated with it or if it contains a pressure datum.

pmdimensions

An ordered list of the dimensions along which the master array is partitioned. Each of these dimensions is one those specified by the nca_dimensions attribute (and is therefore defined as a netCDF dimension in the NCA file).

If missing then it is assumed that the master array comprises a single partition.

Examples
['lat', 'time']
['height']

pmshape

An ordered list containing the number of partitions along each partitioned dimension of the master array. The sizes correspond to the pmdimensions list. This is the shape of the partition matrix.

If missing then a shape of [1] is assumed, i.e. the master array comprises a single partition.

Examples
[2, 3]
[87]

base

A string giving the base for relative file names given by file. May be an absolute or relative URL or location on the local system. If it is a relative location then it assumed to be relative to the local directory or URL base containing the NCA file.

If set then all file names are assumed to have relative paths. If it is an empty string ('') then the base is taken as the local directory or URL base containing the NCA file.

If missing then it is assumed that all file names are absolute paths (local files or URLs).

Examples
'/data/archive'
'../archive/'
'http://archive/files'
''

Partitions

A list whose elements define each of the master array’s partitions. The order of the list is arbitrary since each element contains its (possibly multidimensional) index in the partition matrix.

Examples
[{'index': [0], 'subarray': {'shape': [2, 3], 'ncvar': 'nca_1bE8EBC2c3'}, 'location': [[0, 1], [0, 2]]}]

Each element of the list is an associative array which specifies a partition with the following parameters:

index

An ordered list of indices specifying the position of the partition in the partition matrix. The indices correspond to the pmdimensions list.

Examples
[0]
[2, 1]

Note

Indices count from zero.

location

An ordered list of ranges of indices, one for each dimension of the master data array, which describe the contiguous section of the master data array spanned by this partition. The ranges correspond to the nca_dimensions list.

Each range is a two-element list giving a start and stop index for its dimension. For example, the range [3, 5] is equivalent to indices 3, 4 and 5; and the range [6, 6] is equivalent to index 6.

If the master data array is a scalar then it is an empty list.

Examples
[[2, 2], [3, 5], [2, 56]]
[[0, 0]]
[]

Note

Indices count from zero.

pdimensions (optional)

An ordered list of the partition’s data array dimension names. The specified dimensions are all defined as netCDF dimensions in the NCA file.

If there are any size 1 dimensions of the partition which are not spanned by the master array then the partition’s dimensions must be specified.

If the partition’s data array is a scalar then it may be an empty list.

If missing then it is assumed to be equal to dimensions of the master array.

Examples
['lon', 'time', lat']
[]

Note

If a partition’s data array’s dimensions are not specified and the sub-array is stored in another file then it is required only that the sub-array has the same number of dimensions, with the same physical meaning and in the same order as the master array. For example, if the sub-array were in a different netCDF file, its dimensions may have different names and sizes relative to the equivalent dimensions in the NCA file.

format (optional)

A string naming the format of the file containing the partition’s sub-array.

If missing then the format is assumed to the same as the NCA file.

Examples
'netCDF'
'PP'

pdirections (optional)

An associative array of the partition’s data array dimension directions.

Any dimension not specified is assumed to have the same direction as the corresponding master array dimension. If there are any size 1 dimensions of the partition which are not spanned by the master array then their directions must be specified. The specified dimensions are all defined as netCDF dimensions in the NCA file.

Examples
{'time', true, 'lat': true, 'lon', false}

Note

A size 1 dimension may have an implied direction, e.g. if there are bounds associated with it or if it contains a pressure datum.

units (optional)

A string containing the units of the partition’s data array.

If missing then it is assumed to be equal to units of the master array.

Examples
'm s-1'
''

calendar (optional)

A string containing the calendar of the partition’s data array.

If missing then it is assumed to be equal to calendar of the master array.

Examples
'noleap'

part (optional)

A string defining the part of the sub-array which comprises the partition’s data array.

For each of the partition’s dimensions, the string describes the indices of the sub-array which define the partition’s data array. The indices correspond to to the pdimensions list.

Indices are contained within square or round brackets. Square brackets specify a sequence of indices along that dimension. Round brackets provide a succinct method of describing a strictly monotonic and regularly spaced sequence of indices for the dimension via start, stop and step values. For example, (0, 3, 1) is equivalent to [0, 1, 2, 3] and (10, 4, -2) is equivalent to [10, 8, 6, 4].

If missing or the string '[]' then it is assumed that the partition’s data array is the whole of the sub-array.

Examples
'[(2, 5, 1), [1, 3, 4, 7], (0, 11, 2)]'
'[[5, 2, 1], (5, 1, -4), (0, 0, 1)]'
'[]'

Note

Indices count from zero.

subarray

Parameters required to define the sub-array containing the partition’s data array. Only a subset of these will be required, depending on the storage format of the sub-array.

pshape

An ordered list of the sub-array’s dimension sizes. The sizes correspond to to the pdimensions list.

Examples
[4, 7, 3]

file (optional)

A string naming the file which contains the sub-array. May be a local file or a URL.

If base has been set then the file name is assumed to be relative to it. If base is an empty string ('') then the base is taken as the local directory or URL base containing the NCA file.

If base has not been set then the file name is assumed to be an absolute path (local file or URL).

If missing then it is assumed to be the NCA file itself.

Examples
'/home/me/file.nc'
'../file2.pp'
'file3.nc'
'http://archive/data/file.nc'
'data/file.nc'
''

pdtype (optional)

The data type of the sub-array. Any of the netCDF data type strings are allowed.

If missing then the data type of the master array is assumed.

Examples
'double'
'byte'
'char'

ncvar (optional, but required for netCDF files)

The name of the netCDF variable containing the sub-array.

Examples
'tas'

file_offset (optional)

The non-negative integer word address of the file where the sub-array starts. The start of the file is addressed by zero.

Examples
8460364

lbpack (optional, PP files only)

The PP integer packing code of the sub-array.

If missing then it is assumed to be 0 (unpacked).

Examples
1

scale_factor (optional, non-netCDF files only)

The numeric scale factor (in the CF sense) of the sub-array.

Examples
100.0

For netCDF files, it is assumed that the scale factor will be accounted for when reading the file. Otherwise, if missing then it is assumed to be 1 (unscaled).

add_offset (optional, non-netCDF files only)

The numeric additive offset (in the CF sense) of the sub-array.

Examples
273.15

For netCDF files, it is assumed that the additive factor will be accounted for when reading the file. Otherwise, if missing then it is assumed to be 0 (no additive offset).