Add DataFrame fill_nan/fill_null

[kosiew]

Which issue does this PR close?

Closes #922.

Rationale for this change

DataFusion currently lacks built-in methods for handling missing values (nulls and NaNs) in DataFrames. This functionality is commonly needed in data processing workflows and is available in other DataFrame libraries like pandas and PySpark.

The changes add:

• fill_null() method to replace NULL values with a specified value
• fill_nan() method to replace NaN values with a numeric value in floating-point columns

What changes are included in this PR?

1. Added fill_null() method to DataFrame class:

   • Replaces NULL values with a specified value
   • Handles type casting validation
   • Allows filling specific columns or entire DataFrame
   • Preserves columns where casting fails

2. Added fill_nan() method to DataFrame class:

   • Replaces NaN values with numeric values
   • Only operates on floating-point columns
   • Validates input types and column types
   • Allows filling specific columns or all numeric columns

3. Added comprehensive test cases for both methods covering:

   • Different data types
   • Column subsetting
   • Type validation
   • Error cases

Are there any user-facing changes?

Yes, two new public methods are added to the DataFrame class:

# Fill nulls with a value
df = df.fill_null(0)  # Fill all nulls with 0 where possible
df = df.fill_null("missing", subset=["name"])  # Fill specific columns

# Fill NaN values in numeric columns
df = df.fill_nan(0)  # Fill all NaNs with 0
df = df.fill_nan(99.9, subset=["price"])  # Fill specific columns

[kosiew]

I could not find tests for coalesce which I used for fill_null, so I added these

[timsaucer]

This looks like a very worthwhile and useful addition. Thank you!

We've tried to keep most of the heavier logic on the rust side and to keep the python wrappers as way to convert from rust to pythonic interfaces. Do you think this is a case where doing the logic in the python side makes more sense?

More generally, do you think this is something we can or should upstream to the core datafusion repo? I can assist with that if you like.