how to group timestamp data and filter on it

[Cassa L]

Hi,
I have a data stream (JavaDStream) in following format-
timestamp=second1,  map(key1=value1, key2=value2)
timestamp=second2,map(key1=value3, key2=value4)
timestamp=second2, map(key1=value1, key2=value5)


I want to group data by 'timestamp' first and then filter each RDD for
Key1=value1 or key1=value3 etc.

Each of above row represent POJO in RDD like:
public class Data{
    long timestamp;
    Map<String, String> map;
}

How do do this in spark? I am trying to figure out if I need to use map or
flatMap etc?

Thanks,
LCassa