rnentjes
69fa23a7a9
Relocated MutableByteArrayHandler implementation to correct position in ByteArrayHandler file for better organization. Updated readme to clarify behavior when updating variable-length types like strings for improved compression efficiency.
50 lines
1.8 KiB
Markdown
50 lines
1.8 KiB
Markdown
# typed-byte-arrays
|
|
|
|
This is a small library to make working with byte arrays easier.
|
|
I developed this as an alternative to kotlin serialization which was cumbersome to use and much too slow for my use case.
|
|
|
|
A TypeByteArray is a wrapper around a ByteArray that defines what bytes map to which property. This way there is only data conversion
|
|
to convert the bytes to the property and back when the property is accessed. The data can be sent as the bytearray that it actually is.
|
|
|
|
## Usage
|
|
|
|
This is what a definition could look like:
|
|
|
|
```kotlin
|
|
class Person : TypedByteArray(
|
|
Type("id", DataType.LONG),
|
|
Type("name", DataType.STRING, 100),
|
|
Type("age", DataType.INT),
|
|
Type("rate", DataType.DOUBLE),
|
|
Type("description", DataType.CLOB, 100000)
|
|
) {
|
|
var id by LongProperty("id")
|
|
var name by StringProperty("name")
|
|
var age by IntProperty("age")
|
|
var rate by DoubleProperty("rate")
|
|
var description by ClobProperty("description")
|
|
|
|
constructor(name: String, age: Int) : this() {
|
|
this.name = name
|
|
this.age = age
|
|
this.rate = 1.0
|
|
this.description = "Nothing here yet."
|
|
}
|
|
|
|
constructor(data: ByteArray): this() {
|
|
check(data.size == definition.size) {
|
|
"Invalid data size: ${data.size} != ${definition.size}"
|
|
}
|
|
|
|
this.data = MutableByteArrayHandler(data)
|
|
}
|
|
}
|
|
```
|
|
|
|
The different between STRING and CLOB is the maximum size (32kb vs 2gb). For any type where the length isn't fixed we need to provide max size.
|
|
When such a type is updated, for example we update a string with a shorter string the remaining bytes are zeroed to make compression as efficient as possible.
|
|
|
|
Strings are stored as UTF-8 bytes, there is also a CachedStringProperty which will keep a copy of the string so it isn't converted to a
|
|
native string every time it's accessed.
|
|
|