Parameter Push Down in Python UDF's

[Nintorac]

If I have a query like this

def heavy_function(x):
    return work(x)

c.create_function('heavy_function', heavy_function)
c.query("select id, heavy_function(x) from table limit 10").to_df()

It will compute heavy_function for every row of the table, and then limit to 10 after that. Is this intended behavior or could it be a future optimisation?

---

A bit more hairy of a situation I think but something like this could also be optimised

c.query("select id, heavy_function(x) from table order by rand() limit 10").to_df()