I have four numpy arrays that has about N=5 million rows each.
The shape of them are:

• a: (N, 3, dtype=float32)
• b: (N, 4, dtype=float32)
• c: (N, 3, dtype=uint8)
• d: (N, 4, dtype=uint8)

I want to pack them into a binary file with the same ordering.

This is what I have currently.

buffer = BytesIO()
for row in zip(a, b, c, d):
    buffer.write(row[0].tobytes())
    buffer.write(row[1].tobytes())
    buffer.write(row[2].tobytes())
    buffer.write(row[3].tobytes())

with open(output_path, "wb") as f:
    f.write(buffer.getvalue())

Is there a more efficient method instead of looping it N times like what I had?