Global Attribute for a CDF
Represents a CDF attribute, providing access to the gEntries in a format that looks like a Python list. General list information is available in the python docs: 1, 2, 3.
>>> attribute = cdffile.attrs['attribute_name']
>>> first_gentry = attribute[0]
Each element of the list is a single gEntry of the appropriate type. The index to the elements is the gEntry number.
A gEntry may be either a single string or a 1D array of numerical type. Entries of numerical type (everything but CDF_CHAR and CDF_UCHAR) with a single element are returned as scalars; multiple-element entries are returned as a list. No provision is made for accessing below the entry level; the whole list is returned at once (but Python’s slicing syntax can be used to extract individual items from that list.)
Multi-dimensional slicing is not supported; an entry with multiple elements will have all elements returned (and can thus be sliced itself). Example:
>>> first_three = attribute[5, 0:3] #will fail
>>> first_three = attribute[5][0:3] #first three elements of 5th Entry
gEntries are not necessarily contiguous; a gAttribute may have an entry 0 and entry 2 without an entry 1. len() will return the number of gEntries; use max_idx() to find the highest defined gEntry number and has_entry() to determine if a particular gEntry number exists. Iterating over all entries is also supported:
>>> entrylist = [entry for entry in attribute]
>>> attribute[0:3] = [1, 2, 3]
>>> del attribute[1]
>>> attribute.has_entry(1)
False
>>> attribute.has_entry(2)
True
>>> print attribute[0:3]
[1, None, 3]
Multi-element slices over nonexistent gEntries will return None where no entry exists. Single-element indices for nonexistent gEntries will raise IndexError. Assigning None to a gEntry will delete it.
When assigning to a gEntry, the type is chosen to match the data; subject to that constraint, it will try to match (in order):
- existing gEntry of the same number in this gAttribute
- other gEntries in this gAttribute
- data-matching constraints described in CDF.new().
See also