Whoops! I forgot there was a bug with split which was just fixed last week:
I was on main locally so I didn’t see it. To workaround, you’ll need to do a pipeline like this:
df
|> DF.put("text", S.split(df["text"], ", "))
|> DF.explode("text")
(@bdarla’s approach works too.)
Or you can use main until the next release. Sorry about that!
