Composite key Issues

[karthickkrd21]

I am using spark hbase connector version 1.1.1-2.1-s_2.11 to read and write data into hbase with spark 2.1. I am quite new to this, few questions on the usage

1. how does the predicate pushdown works here for the composite key, I went through one of the issue where it said works only on the first dimension of the composite key. is it same now or any improvement being done here for all part of the composite key.
   2.For non composite key how does the predicate push down happen for a rowkey and columns,
2. Please suggest which one is better composite/non composite key.
3. Please share some details on partition pruning