String data points are read wrong.

Querying string data points returns ByteBuffer#toString instead of converting the bytes to a UTF8 string. The write operation is done with the DataOutput#writeUTF so the read should be done with DataInput#readUTF these methods generate bytes in a custom format. So using the String constructor would not work. I think fixing the read instead of the write makes more sense because of backwards compatability?

The read operation is done wrongly here. And you can see the write operation here.

Example:

[brianhks]

Now I got it.

[brianhks]

Schema was wrong and didn’t match the thrift schema. I just fixed it and checked it into develop.