Python: NumPy Boolean Data Type

I have been learning Python - more specifically NumPy and Pandas, and came across this.

NumPy stores the Boolean Data Types (True/False) as 8 bits or 1 byte.

As a data engineer, my immediate reaction was, "That is a wasteful allocation of resources. Why did they do this?" Surely they could achieve this with 1 bit?

Well, if you look a bit deeper, there is a very good reason.

NumPy optimizes for speed, not minimal memory.

NumPy uses 1 byte per boolean because:

• CPU architecture CPUs work with bytes, not individual bits.
• Vectorized operations require byte/word alignment.
• NumPy arrays assume fixed-size, addressable elements.
• Slicing/indexing would be complicated with bit-level addressing.
• Interop with C, Fortran, SIMD, BLAS all expect byte-sized elements.

So yes: it uses more memory, but computation is dramatically faster.