Basically, I have a data set where a document can contain an array of variable length. I'm writing a codec, so all I have is BsonReader. The problem is readStartArray() returns void rather than returning, for example, the number of elements. Also, I don't see any methods that allow one to test when the end of the array has been reached (apart from trying readEndArray() and catching the exception).
Any ideas?
I found the documentation for custom serialization to be lacking too, but you just need to work with the Reader.State property. Once you've identified the start of the array you need to iterate over each of the elements until state == BsonReaderState.EndOfArray. You need to ensure that you only call ReadBsonType when the state is "type".
The following example deserializes an array of documents:
context.Reader.ReadStartArray();
bool ctn = true;
do
{
//If in "type" state we can call ReadBsonType. The order is important.
if (context.Reader.State == BsonReaderState.Type)
{
context.Reader.ReadBsonType();
}
//Now that we're in a value state we can read the value safely
if (context.Reader.State == BsonReaderState.Value)
{
if (context.Reader.CurrentBsonType == BsonType.Document)
{
//Handle the document here. In this case pass I it to a private method
DeserializeChild(context, args, node);
}
//state will now be "type", so continue do loop and recheck type
continue;
}
if ( BsonReaderState.EndOfArray == BsonReaderState.EndOfArray)
{
context.Reader.ReadEndArray();
ctn = false;
}
} while (ctn);
ReadBsonType() when the state of the reader is context.Reader.State == BsonReaderState.Type, that the CurrentBsonType and State properties of the reader are stale and invalid.If you're using java, then you need to use BsonArrayCodec. Although I think something similar can be found in client libraries for other programming languages too.
For example:
List<String> names = new ArrayList<>();
BsonArray underlyingBsonArray = bsonArrayCodec.decode(reader, context);
for (BsonValue bsonValue : underlyingBsonArray) {
names.add(bsonValue.asDocument().getString("name").getValue());
}
Ah, it's done with:
while (reader.readBsonType() == BsonType.DOCUMENT) {
reader.readStartDocument();
// ...
reader.readEndDocument();
}
(it's an array of documents in this case, I'm still new to BSON so I don't know how it would work for an array of values).