Add support for encoding and decoding Decimals #93
Conversation
lib/avro_ex/encode.ex
Outdated
| is_number(value) -> | ||
| value / :math.pow(10, -scale) | ||
|
|
||
| match?(%Decimal{}, value) -> |
There was a problem hiding this comment.
how does this work with optional dependencies? Will this compile if the user does not have :decimal installed?
There was a problem hiding this comment.
Indeed, it doesn't. Thanks !
I have fixed it by trying to follow the same approach followed in Jason.
I've tested it in an empty project, it seems to work.
| @@ -316,6 +316,34 @@ defmodule AvroEx.Decode.Test do | |||
|
|
|||
| assert Time.truncate(time, :millisecond) == now | |||
| end | |||
|
|
|||
| test "decimal" do | |||
| schema = "test/fixtures/decimal.avsc" |> File.read!() |> AvroEx.decode_schema!() | |||
There was a problem hiding this comment.
it would be useful to test if round tripping the value through the encoder results in the same result
There was a problem hiding this comment.
Indeed. There are now some round-trip tests in the encode_test. Do you think it is sufficient our should I add some here as well ?
| if :math.pow(2, bits) > abs(value) do | ||
| bits | ||
| else | ||
| value_size(value, bits + 8) |
There was a problem hiding this comment.
are you sure that this will terminate for all input values?
There was a problem hiding this comment.
Hum, I am not. Should we stop when we reach 64 bytes ?
There was a problem hiding this comment.
On second thought: I don't actually see a situation where this won't terminate. The value of bits can only by positive multiples of 8, so we're always bound to eventually get a 2^bits that's bigger than any arbitrary value.
WDYT ?
|
@davydog187 is there anything I can do to help with this ? |
|
@jbruggem looks good to me! Can you please fix dialyzer and the tests? Then we can get this released |
|
@davydog187 thanks ! The compilation error is fixed. |
|
Thank you! |
This PR proposes to add support for Decimal values in
avro_ex.As decimal values aren't accurately represented by native number in Elixir, it proposes to (optionally) decode these decimals losslessly as Decimal structs.
In order to ensure consistency with the official implementation, this PR proposes to test the decoding/encoding using a reference Avro encoded file generated with a Java implementation.