From Slack @ruebot noted that if you say

import io.archivesunleashed._
import io.archivesunleashed.df._
val df = RecordLoader.loadArchives("/tuna1/scratch/nruest/geocites/warcs/9/image_test/*", sc).extractImageDetailsDF();
val res = df.select($"bytes").orderBy(desc("bytes")).saveToDisk("bytes", "/tuna1/scratch/nruest/geocites/images/")

Then in the images directory you get a bunch of files with the preceding -.

We should refine and then document this example so that it is say

val res = df.select($"bytes").orderBy(desc("bytes")).saveToDisk("bytes", "/tuna1/scratch/nruest/geocites/images/geo")

where each image would then be geo-HASH.gif etc.

Opening an issue just so I don't forget to do this.