This is a draft
The following Pull Request:

• Closes There are four ways to specify Attributes on DAP4 spec - Pydap should be able to parse all #406. Correctly parses an attribute with multiple values to generate a list of values.
• Tests are added to demonstrate the new or fixed behavior, and all tests pass
  on a local environment.

example

This is included in the testing

dmr = """
<Dataset name="foo">\n
<Int32 name="x">\n
    <Attribute name="attr" type="Float32">\n
        <Value></Value>\n
        <Value>2</Value>\n
        <Value value="1"/>\n</Attribute>\n
</Int32>\n
</Dataset>
"""
dom_et = DMRParser(dmr).node
dataset = DMRParser(dmr).init_dataset()
variables = get_variables(dom_et)
get_attributes(variables['x']['element'])['attr']

the last line returns the list for attribute 'attr'. All values, excep None, are Float32, associated with variable 'x' of type int32

[None, 2.0, 1.0]