Elixir parse binary data?

Question

for example:

I have a binary look like this:

 bin1 = "2\nok\n3\nbcd\n\n"

or

 bin2 = "2\nok\n3\nbcd\n1\na\n\n"

and so on...

The format is

 byte_size  \n  bytes \n byte_size  \n  bytes \n  \n

I want parse binary get

  ["ok", "bcd"]

how to implement in Elixir or Erlang ?

Go version

a Go version parse this

func (c *Client) parse() []string {
    resp := []string{}
    buf := c.recv_buf.Bytes()
    var idx, offset int
    idx = 0
    offset = 0

    for {
        idx = bytes.IndexByte(buf[offset:], '\n')
        if idx == -1 {
            break
        }
        p := buf[offset : offset+idx]
        offset += idx + 1
        //fmt.Printf("> [%s]\n", p);
        if len(p) == 0 || (len(p) == 1 && p[0] == '\r') {
            if len(resp) == 0 {
                continue
            } else {
                c.recv_buf.Next(offset)
                return resp
            }
        }

        size, err := strconv.Atoi(string(p))
        if err != nil || size < 0 {
            return nil
        }
        if offset+size >= c.recv_buf.Len() {
            break
        }

        v := buf[offset : offset+size]
        resp = append(resp, string(v))
        offset += size + 1
    }

    return []string{}
}

Thanks

bitwalker · Accepted Answer · 2016-06-28 22:24:31Z

8

A more flexible solution:

result = bin 
|> String.split("\n") 
|> Stream.chunk(2)
|> Stream.map(&parse_bytes/1)
|> Enum.filter(fn s -> s != "" end)

def parse_bytes(["", ""]), do: ""
def parse_bytes([byte_size, bytes]) do
  byte_size_int = byte_size |> String.to_integer
  <<parsed :: binary-size(byte_size_int)>> = bytes
  parsed
end

edited Jun 28, 2016 at 22:24

answered Jun 26, 2014 at 19:34

bitwalker

9,2891 gold badge39 silver badges29 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

lidashuang Over a year ago

thanks, I add a version to description this question

jisaacstone Over a year ago

Note that this solution is not good for parsing a large amount of data because it will iterate over several times. The 2nd solution in lidashuang's answer will avoid that issue.

bitwalker Over a year ago

I've updated mine to use the Stream functions instead of Enum. This will sidestep the multiple iteration issue as well.

Jonas Adler Over a year ago

note that in newer elixir versions you will have to write binary-size(byte_size_int) instead of [binary, size(byte_size_int)]

Eric Platon · Accepted Answer · 2016-04-05 04:27:58Z

4

I wrote a solution:

  defp parse("\n") do
    []
  end

  defp parse(data) do
    {offset, _} = :binary.match(data, "\n")
    size = String.to_integer(binary_part(data, 0, offset))
    value = binary_part(data, offset + 1, size)

    len = offset + 1 + size + 1
    [value] ++ parse(binary_part(data, len, byte_size(data) - len))
  end

The Elixir mailing list provides another one:

  defp parse_binary("\n"), do: []

  defp parse_binary(binary) do
    {size, "\n" <> rest} = Integer.parse(binary)
    <<chunk :: [binary, size(size)], "\n", rest :: binary>> = rest
    [chunk|parse_binary(rest)]
  end

edited Apr 5, 2016 at 4:27

Eric Platon

10.2k6 gold badges43 silver badges50 bronze badges

answered Jun 27, 2014 at 10:37

lidashuang

2,0034 gold badges21 silver badges23 bronze badges

Collectives™ on Stack Overflow

Elixir parse binary data?

Go version

2 Answers 2

4 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

Go version

2 Answers 2

4 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related