Transform Values with a Comprehension

A loop can build a new list while keeping every step visible. After that loop is understood, a list comprehension can express the same simple rule in one place.

Begin with the Complete Loop

Suppose a sensor needs a correction of 0.5 degrees:

The result is [18.5, 21.5, 20.5, 23.5]. The source list is unchanged; the loop builds a separate result list.

The equivalent list comprehension is:

corrected = [reading + 0.5 for reading in readings]

Read it from the for clause:

1. take each reading from readings;
2. calculate reading + 0.5;
3. place that result in the new list.

[reading - 1 for reading in [18, 21, 20]]

Add One Filter When Some Items Should Be Omitted

Chapter 3 used None for an unavailable reading. The loop form filters those values before transforming the others:

The matching comprehension places the filter after the for clause:

Read the filter first: only an available reading continues to the result expression. A filter may make the output shorter than the input.

This is different from a conditional expression, which produces one result for every input:

labels = ["high" if reading >= 22 else "ordinary" for reading in [18, 22, 20]]

The three inputs produce three labels. The conditional expression chooses which label to add; it does not omit an item.

[reading * 2 for reading in [3, None, 5] if reading is not None]

Expand a Nested Form Before Trusting It

A small nested comprehension can flatten two short runs of readings:

Expand it into the loops that execute in the same order:

Both forms build [18, 20, 21, 23]. The explicit loop is usually clearer when there are several filters, several calculations, or intermediate values worth naming. Compactness is not a reason to hide a procedure.