← All lessons

Assign and fill

Write into slices, fill, and np.where as an expression.

You can write into a slice (a view) or build a new array with np.where. In-place updates are fast; expressions are easier to reread.

Goal

Overwrite a row or a mask, fill a block, and use np.where as an expression.

Write a row

# Rows: Nairobi, Mombasa, Kisumu  ·  cols: A, B, C
units = np.array(
    [
        [12, 7, 2],
        [9, 6, 0],
        [3, 11, 1],
    ],
    dtype=float,
)
units[2] = [4, 4, 4]
print(units)

Kisumu is now 4, 4, 4. That assignment used the view of row 2.

Write where a mask is True

units = np.array(
    [
        [12, 7, 2],
        [9, 6, 0],
        [3, 11, 1],
    ],
    dtype=float,
)
units[units == 0] = np.nan
print(units)

Integer arrays cannot hold nan — this grid is float, so the zeros become missing. The Types chapter covers the cast.

np.where as an expression

units = np.array(
    [
        [12, 7, 2],
        [9, 6, 0],
        [3, 11, 1],
    ],
    dtype=float,
)
out = np.where(units >= 9, units, 0)
print(out)

Busy cells keep their units; the rest become 0. units itself is unchanged.

fill

buf = np.empty((2, 3))
buf.fill(10.5)
print(buf)

empty does not zero the buffer — fill (or zeros / full) makes the contents defined.

Tip

Prefer out = np.where(...) when you want to keep the original. Prefer arr[mask] = value when you mean to edit in place.