Extends parquet fuzz tests to also tests nulls, dictionaries and row groups with multiple pages (#1053)

[tustvold]

Which issue does this PR close?

Closes #1053.

Rationale for this change

See ticket

What changes are included in this PR?

This extends the parquet fuzz tests to also tests nulls, dictionaries and row groups with multiple pages. Currently this runs into what appears to be a bug in the null handling for ArrowArrayReader. This is likely the same as in apache/datafusion#1441 - I have temporarily switched back to ComplexObjectArrayReader to get the test to pass, and will look into a fix prior to marking this ready for review. This has been fixed by #1130

Are there any user-facing changes?

No, this only adds tests

[tustvold]

This didn't seem to serve a purpose, as it was always set in such a way as to read all the data, so I removed it

[alamb]

I agree that it is redundant when record_batch_size is provided (which means the data is not all read in one big chunk, but is read in record_batch_size chunks)

[codecov-commenter]

Codecov Report

Merging #1110 (ec79c43) into master (719096b) will increase coverage by 0.01%.
The diff coverage is 94.16%.

❗ Current head ec79c43 differs from pull request most recent head 87ea9a1. Consider uploading reports for the commit 87ea9a1 to get more accurate results

@@            Coverage Diff             @@
##           master    #1110      +/-   ##
==========================================
+ Coverage   82.55%   82.56%   +0.01%     
==========================================
  Files         169      169              
  Lines       50456    50535      +79     
==========================================
+ Hits        41655    41726      +71     
- Misses       8801     8809       +8     

Impacted Files	Coverage Δ
parquet/src/arrow/arrow_reader.rs	89.93% <94.06%> (+0.61%)	⬆️
parquet/src/util/test_common/rand_gen.rs	82.69% <100.00%> (+0.87%)	⬆️
arrow/src/datatypes/field.rs	53.79% <0.00%> (-0.31%)	⬇️
parquet_derive/src/parquet_field.rs	66.21% <0.00%> (-0.23%)	⬇️
arrow/src/array/transform/mod.rs	85.56% <0.00%> (-0.14%)	⬇️
parquet/src/arrow/converter.rs	69.56% <0.00%> (+0.86%)	⬆️

---

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 719096b...87ea9a1. Read the comment docs.

[tustvold]

Thanks to @yordan-pavlov 's work on #1130 this now passes on master 🎉

[alamb]

LGTM

Nice work @tustvold

[alamb]

I agree that it is redundant when record_batch_size is provided (which means the data is not all read in one big chunk, but is read in record_batch_size chunks)