Skip to content

numpy/nan-to-int

Rule nan-to-int
Pack NumPy
Severity warning
Fix none, edit by hand
Library versions every version

Casting an array that holds NaN to an int type turns each NaN into -9223372036854775808. Keep the float array, or replace NaN with a whole number before the cast.

counts = np.where(seen, totals, np.nan).astype(int)
counts = np.where(seen, totals, 0).astype(int)

The rule reports np.where(m, x, np.nan).astype(int), which turns each NaN into -9223372036854775808.

xpd has no automatic fix, so make the change by hand.

xpd prints this message with the issue. A name that starts with $ stands for code from your file.

This casts an array that holds np.nan to $T, and NaN becomes -9223372036854775808. Keep the float array, or replace NaN with a whole number before the cast.

Browse the other rules of the NumPy pack.