Choose an observed middle value with median_low or median_high
A middle value used as a discrete setting may need to come from the observed settings. Averaging the two middle observations can produce a value that was never used.
Run the example
Tested on Linux with Python 3.12.14. No third-party packages are required. In one Bash session, create a disposable directory:
start_dir="$PWD"
work_dir="$(mktemp -d)" || exit 1
cd -- "$work_dir" || exit 1Save this as demo_python_median_low_high_observed_values.py. Use that filename to avoid shadowing a standard-library module.
from statistics import median, median_low, median_high, StatisticsError
for settings in ([1, 2, 8, 16], [1, 2, 8]):
print(f"{settings}: median={median(settings)}, low={median_low(settings)}, high={median_high(settings)}")
try:
median_low([])
except StatisticsError:
print("empty settings: rejected")Run it:
python3 demo_python_median_low_high_observed_values.pyOutput:
[1, 2, 8, 16]: median=5.0, low=2, high=8
[1, 2, 8]: median=2, low=2, high=2
empty settings: rejectedWith an even number of observations, the low and high variants choose the lower and upper middle observations respectively. With an odd count they agree. Here the ordinary median of the even-sized sample is five, while the observed middle settings are two and eight.
Input policy and limits
Choosing the lower or upper middle observation is an application policy, not a guarantee about capacity or safety. Validate allowed settings separately, handle missing observations, and reject or remove NaNs before ordering numeric data. These functions describe the middle of the observations; they do not find the most frequent setting.
Clean up
In the same Bash session, remove the source file and the empty working directory:
rm -- demo_python_median_low_high_observed_values.py
cd -- "$start_dir" || exit 1
rmdir -- "$work_dir"
unset start_dir work_dirIf the directory is not empty, inspect its contents before deleting anything else.
References
Acknowledgements
This post was created with assistance from AI.