In my efforts to solve my issue here, I'm using ArangoDb's AQL language to split a collection into two. As creating graphs can only be done on "normalized" collections. But my query is eating up all of my RAM (16gb).
FOR comment IN collection_name
COLLECT author = comment.author
, subreddit = comment.subreddit
WITH COUNT INTO counter
RETURN {author, counter, subreddit}
Is there a way I could dump the results directly to disk without pulling it into RAM? Or a way I could split it directly without using a query?