apache spark - How to cast Any Scala to Double -

i iterating dataframe , need apply custom code each row i'm doing

convertedpaths.foreach { row : row =>    cfmap = row.getvaluesmap(channelset.toseq) 

assume channelset set of column names. i've declared cfmap of type [string, any] getvaluesmap (as understood return data type of column)

also, columns of long type trying :

channelset.foreach { key : string =>    var frequency = cfmap.get(key).get.asinstanceof[double]    var value = c * frequency  

given c variable of type double , value needs product of c , frequency gives me following error :

overloaded method value * alternatives: (x: double)double <and> (x: float)double <and> (x: long)double <and> (x: int)double <and> (x: char)double <and> (x: short)double <and> (x: byte)double cannot applied (any) 

why asinstanceof[double] not correct solution , solution this?

although doesnt answer problem statement , i'm trying read on accumulators on how collect result major portion of logic can run inside convertedpaths.foreach loop.

i able resolve problem assigning

var cfmap = map[string, long]()  

instead of

var cfmap = map[string, any]()  

as know columns in convertedpaths long type