With Rust Polars, is there a good way to replace a specified value in a series like fill_null?

Question

I have a dataset where null/missing values are represented by 0. So I'd like to do something like c.replace_val(0, "forward"). What is a good/easy/efficient way to do that? Thanks.

Answer 1

1. if you are reading from a file you can specify your null_values and then use a .forward_fill() in one pass:

   d = pl.read_csv('file.csv', null_values=0)

   d.a.fill_null('forward')

2. if your are not reading from a file, I am afraid you need to impute your 0s first and then replace them. You can chain when/then/otherwise/fill_null. E.g.,

   d.with_column( pl.when(col("a") == 0) .then(None) .otherwise(col("a")) .fill_null("forward")