statistics: mean() Can't Mix Decimal With Other Types (tested)

statistics.mean() is documented as working across int, float, Decimal, and Fraction data. We measured what happens when a single call mixes two of those types together in the same list, and found it depends heavily on which two types -- some combinations work transparently, others raise a distinctive error that doesn't look like Python's usual arithmetic TypeError.

1. Mixing Fraction and float works fine and returns float

import statistics as st
from fractions import Fraction

result = st.mean([Fraction(1, 2), 0.5])
print(result, type(result))

0.5 <class 'float'>

Measured: statistics.mean() has its own internal type-coercion logic, separate from Python's normal operator dispatch -- it successfully mixes Fraction and float values in one call, promoting the result to float.

2. Mixing Decimal with float or Fraction raises TypeError with an unusual message

import statistics as st
from fractions import Fraction
from decimal import Decimal

try:
    st.mean([1.5, Decimal("0.5")])
except TypeError as e:
    print("TypeError:", e)

try:
    st.mean([Fraction(1, 2), Decimal("0.5")])
except TypeError as e:
    print("TypeError:", e)

TypeError: don't know how to coerce float and Decimal
TypeError: don't know how to coerce Decimal and Fraction

Measured: both combinations raise TypeError, but with the message "don't know how to coerce <type> and <type>" -- this comes from statistics's own internal _coerce() helper, not from Python's normal +// operators, so it reads differently from the TypeError you'd get calling 1.5 + Decimal("0.5") directly ("unsupported operand type(s) for +"). Grepping logs for the standard arithmetic error text won't catch this one.

3. Decimal only coerces cleanly with int, not with float or Fraction

import statistics as st
from decimal import Decimal

print(st.mean([1, 2, Decimal("3")]))
print(type(st.mean([1, 2, Decimal("3")])))

2
<class 'decimal.Decimal'>

Measured: Decimal mixed with plain int works fine and the result stays a Decimal -- it's specifically float and Fraction that Decimal can't be coerced with inside statistics, matching the fact that plain Python arithmetic rejects Decimal+float too (though, as shown above, Fraction+float arithmetic works directly while Fraction+Decimal arithmetic doesn't -- yet both raise inside statistics.mean() once Decimal is one of the two types). A data pipeline that stores measurements as plain float in one place and switches to Decimal for a "more precise" column elsewhere will crash the first time both land in the same statistics.mean() call, with an error message that doesn't mention "Decimal" or "float" in the way a normal arithmetic error would make obvious at a glance -- read the full message, or convert everything to one exact type (e.g. Decimal via each value's str(), not float()) before calling statistics functions on mixed-provenance data.

Written with AI assistance. Measured on 2026-10-04 with Python 3.13.5's built-in statistics module: called mean() on lists mixing int, float, Fraction, and Decimal values in different combinations, and compared the exact error to Python's normal arithmetic TypeError. PlainHeaders is our extension; Python's statistics, fractions, and decimal modules are built in.