Carrying over values from most recent row satisfying a condition

Question

I have a pyspark dataframe that looks like (a massively larger version of) the following:

+---+---+----+----+
| id|  t|type| val|
+---+---+----+----+
|100|  1