I can load multiple csv files by doing something like:

paths = ["file_1", "file_2", "file_3"]
df = sqlContext.read
       .format("com.databricks.spark.csv")
       .option("header", "true")
       .load(paths)

But this doesn't seem to preserve the order in |paths|.

In particular, I'm trying to have a monotonically increasing id that spans the data in all files.