About numbers, binaries and bytes



How can I get the size (bits or bytes) of a variable?

I don’t know if the next are correct:

  • 0x002A 4B (if integers are stored in 4B)
  • 0x2A00 4B (if integers are stored in 4B)
  • << 0x002A >> 2B
  • << 0x2A00::size(16) >> 2B
  • << 0x2A00::size(8) >> 1B (<< 0x00 >>)


And, also, Why bytes-size and size are not compatible?

@signature << 0xFFD8::size(16) >>

Changed by:

@signature << 0xFFD8::bytes-size(2) >>

Gives me this error:

** (CompileError) lib/file.ex:13: conflicting type specification for bit field: "binary" and "integer"
    (elixir) src/elixir_bitstring.erl:87: :elixir_bitstring.expand_specs/5
    (elixir) src/elixir_bitstring.erl:29: :elixir_bitstring.expand/6
    (elixir) src/elixir_bitstring.erl:12: :elixir_bitstring.expand/4
    (stdlib) lists.erl:1354: :lists.mapfoldl/3
    (stdlib) lists.erl:1355: :lists.mapfoldl/3
    (elixir) expanding macro: Kernel.@/1


1 Like

You can learn about the memory consumption of various erlang base types in the erlang documentation, chapter “advanced”.

<< 0xFFD8::bytes-size(2) >> doesn’t work, because the integer literal already specifies a unit of 1, you can’t override it.

I don’t understand why it allows overriding the unit of 1 (byte?) with 16 bits, but not 2 bytes.


Integer literals imply a unit(1), the unit can’t be overwritten once set. Also you can’t change the size once its set.

unit specifies the number of bits to use for a single “fragment”, size specifies the number of “fragments”.

I think what confuses you is bytes-size. It is just an alias for binary-size. If you want to specify the integer size with bytes instead of bits, use:

<< 0xFFD8::size(2)-unit(8) >>

Hey cool, I didn’t know that you can specify a new unit even on literals. Thank you!

1 Like

Thank you very much to you two @NobbZ and @grych.