ByteStringArrayCodecs

CBOR has a very compact encoding for byte arrays because binary data are “first-class”.
While JSON itself has no such built-in support for binary representation Base Encodings can provide an often acceptable work-around for the problem. (See also this chapter for more information about borer’s out-of-the-box support for Base Encodings!)

Sometimes, however, the need arises to encode larger amounts of data as efficiently as binary data, even if they are not immediately represented as byte arrays, for example arrays of primitive numbers such as:

• Array[Short]
• Array[Int]
• Array[Long]
• Array[Float]
• Array[Double]

By default borer will encode these types like any other array, which can add quite a bit of overhead, especially in the case of JSON.

In order to allow for more efficient encodings (as well as faster encoding and decoding processes) borer provides special codecs for these array types that you can “enable” with one of these two imports:

• import io.bullet.borer.ByteStringArrayCodecs.BigEndian.*
• import io.bullet.borer.ByteStringArrayCodecs.LittleEndian.*

With one of these imports in place an Array[Int] for example will be encoded as a raw byte array, with each Int value being represented as four bytes (in the respective byte order). Conversion to and from byte arrays is done as efficiently as the platform allows.

Also, the ByteStringArrayCodecs are compatible with custom byte array encoders (like the Base Encoders).