This attribute provides a human readable description of the content of the variable. It should make clear what the source of the data was (e.g. measured, climatology, derived, etcetera).
This attribute is only applicable for HDF4 files (netCDF-3 uses named dimensions and HDF5 uses dimension scales). This attribute stores the type of each dimension of the associated variable as a comma-separated list of dimension type names. The number of dimension types should equal the number of dimensions of the variable.
This attribute is used for data that has a physical unit. It should provide the unit in a form compatible with the
unitsattribute is expected to be available for any variable defining a quantity. If a variable represents a dimensionless quantity the
unitsstring should be an empty string (or have the value
1in case empty strings are not supported). The plural attribute name
unitsis only used when storing HARP data in a file (for compatibility reasons with other standards such as netCDF-CF); inside data structures of HARP variables within C, Python, etc. the (more accurate) singular term
valid_min[int8, int16, int32, float, double] (optional)
Provides the minimum value below which the data is to be considered invalid. The data type of this attribute should match the data type of the associated variable. This attribute is not allowed to be present for variables of type string. For variables of numeric type, this attribute should only be present if the variable actually contains values below this threshold that are to be interpreted as missing or invalid values.
valid_max[int8, int16, int32, float, double] (optional)
Provides the maximum value above which the data is to be considered invalid. The data type of this attribute should match the data type of the associated variable. This attribute is not allowed to be present for variables of type string. For variables of numeric type, this attribute should only be present if the variable actually contains values above this threshold that are to be interpreted as missing or invalid values.
Provides the list of string values of the enumeration labels of a categorical variable. For the in-memory representation (C, Python, etc.) this attribute is called
enum_nameand consists of a list of strings. Within a netCDF-3, HDF4, or HDF5 file this attribute is stored as a single string attribute named
flag_meaningswhich contains a space separated list of the label names.
_FillValue is not used by HARP. Whether a value is valid is purely determined by the
Note that files using the HARP data format can include variable attributes in addition to the ones mentioned above
(such as e.g. a
_FillValue attribute). However, such attributes will be ignored when the product is imported.