fix 305 by using a scalar uint as param for zero param functions

[jimexist]

Which issue does this PR close?

fix 305 by using a null array as param for zero param functions

Closes #305.

Rationale for this change

Why are you proposing this change? If this is already explained clearly in the issue then this section is not needed.
Explaining clearly why changes are proposed helps reviewers understand your changes and offer better suggestions for fixes.

What changes are included in this PR?

There is no need to duplicate the description in the issue here but it is sometimes worth providing a summary of the individual changes in this PR.

Are there any user-facing changes?

If there are user-facing changes then we may require documentation to be updated before approving the PR.

If there are any breaking changes to public APIs, please add the breaking change label.

[codecov-commenter]

Codecov Report

Merging #307 (9f93d39) into master (63eaeee) will decrease coverage by 0.01%.
The diff coverage is 44.44%.

@@            Coverage Diff             @@
##           master     #307      +/-   ##
==========================================
- Coverage   76.19%   76.18%   -0.02%     
==========================================
  Files         141      141              
  Lines       23786    23792       +6     
==========================================
+ Hits        18123    18125       +2     
- Misses       5663     5667       +4     

Impacted Files	Coverage Δ
datafusion/src/physical_plan/functions.rs	89.18% <44.44%> (-0.24%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 63eaeee...9f93d39. Read the comment docs.

[alamb]

I think this code is ok. It would be nice to have unit tests, but I think #303 will add sufficient converage.

It would be nice to add some additional documentation about this as it is "not obvious" to a casual reader.

Thank you @jimexist

[alamb]

I think this approach, while a hack, seems reasonable to me

I agree with the comments on #303 (comment) that this does seem like a hack (it would be nicer to have an enum or something that made this case explicit).

I would like to see this expectation documented near the place where a user would define a UDF as well - either as an example or as a doccomment. Perhaps somewhere in
https://github.com/apache/arrow-datafusion/blob/master/datafusion/src/physical_plan/udf.rs

Or bonus points for adding an example:
https://github.com/apache/arrow-datafusion/blob/master/datafusion-examples/examples/simple_udf.rs

cc; @msathis who I think faced a similar challenge in now(): #288

[jimexist]

@alamb tried a different approach using TryFrom trait, so it's more obvious. the call site convention I'll try to code in #303

[jimexist]

Docs in the next PR is fine. I think @jorgecarleitao had an alternate suggestion in #303 (comment)

see #328

[jimexist]

I think this code is ok. It would be nice to have unit tests, but I think #303 will add sufficient converage.

It would be nice to add some additional documentation about this as it is "not obvious" to a casual reader.

Thank you @jimexist

I'll try to add comments in the next pull request - is that okay?

[alamb]

Docs in the next PR is fine. I think @jorgecarleitao had an alternate suggestion in #303 (comment)

[jimexist]

if #328 can be merged then this can be closed

[jimexist]

closing this as we are aligned on using NullArray.