Additional Tuple Techniques
Using an immutable namedtuple as a record
In Chapter 3, Functions, Iterators, and Generators, we showed two common techniques
to work with tuples. We've also hinted at a third way to handle complex structures.
We can do any of the following, depending on the circumstances:
- Use lambdas (or functions) to select a named item using the index
- Use lambdas (or functions) with *parameter to select an item by parameter
name, which maps to an index - Use namedtuples to select an item by attribute name or index
Our trip data, introduced in Chapter 4, Working with Collections, has a rather complex
structure. The data started as an ordinary time series of position reports. To compute
the distances covered, we transposed the data into a sequence of legs with a start
position, end position, and distance as a nested three-tuple.
Each item in the sequence of legs looks as follows as a three-tuple:
first_leg= ((37.54901619777347, -76.33029518659048), (37.840832,
-76.273834), 17.7246)
This is a short trip between two points on the Chesapeake Bay.
A nested tuple of tuples can be rather difficult to read; for example, expressions such
as first_leg[0][0] aren't very informative.
Let's look at the three alternatives for selected values out of a tuple. The first technique
involves defining some simple selection functions that can pick items from a tuple by
index position:
start= lambda leg: leg[0]
end= lambda leg: leg[1]
distance= lambda leg: leg[2]
latitude= lambda pt: pt[0]
longitude= lambda pt: pt[1]
With these definitions, we can use latitude(start(first_leg)) to refer to a
specific piece of data.