java.lang.ClassCastException on shortestPaths calculation

[cj2001]

I have a graph, g, which is roughly 4.5M vertices and 43M edges as:

scala> g res10: org.graphframes.GraphFrame = GraphFrame(v:[id: int], e:[src: int, dst: int])

I would like to calculate the shortest path between two given vertices, so I have run

val shortest_path = g.shortestPaths.landmarks(Seq("1","2")).run().

based on the documentation provided here. While this runs and generates the shortest_path dataframe, when I try to .show() that dataframe I get:

java.lang.ClassCastException: java.lang.Long cannot be cast to java.lang.Integer.

Suggestions?

[felixcheung]

It's been a while but I recall some issue with shortestpaths - it looked like scala unbox error like this

java.lang.ClassCastException: java.lang.Long cannot be cast to java.lang.Integer
at scala.runtime.BoxesRunTime.unboxToInt(BoxesRunTime.java:106)

(would be good if anyone would like to help fix)

I was able to workaround with id etc casted to string

v1 = v.select(v.id.cast("string").alias("id"))
e1 = e.select(e.src.cast("string").alias("src"), e.dst.cast("string").alias("dst"), e.relationship)
g1 = GraphFrame(v1, e1)

[cj2001]

This appears to have worked. Thank you! (I leave it up to you whether you want to keep this issue open as a reminder to look into that error or not.)

[felixcheung]

let's see if we can fix this...

[marty90]

+1