Added support for remaining non-nested datatypes

[jorgecarleitao]

This PR adds support to dictionary-encoding (parquet) to the remaining non-nested datatypes.

This is demonstrated by:

1. a roundtrip test of all columns in the generated_dictionary IPC file pair (containing dictionaries with different types and validities)
2. a rust -> pyspark test (equality over all values)

Together, they demonstrate that

• spark reads dictionary-encoded arrays written by this crate
• this crate reads dictionary-encoded arrays written by this crate

where "read" is in the sense that data integrity is preserved over all values (nulls or not).

Tests against pyarrow currently fail, likely due to ARROW-13486 and ARROW-13487.

[codecov]

Codecov Report

Merging #336 (b7e2352) into main (8e96ec4) will increase coverage by 0.01%.
The diff coverage is 66.03%.

@@            Coverage Diff             @@
##             main     #336      +/-   ##
==========================================
+ Coverage   80.96%   80.97%   +0.01%     
==========================================
  Files         325      326       +1     
  Lines       21077    21167      +90     
==========================================
+ Hits        17066    17141      +75     
- Misses       4011     4026      +15     

Impacted Files	Coverage Δ
src/io/parquet/read/mod.rs	51.21% <30.55%> (-18.79%)	⬇️
src/io/parquet/read/binary/dictionary.rs	80.00% <80.00%> (ø)
src/io/parquet/write/dictionary.rs	72.72% <100.00%> (+29.48%)	⬆️
src/io/parquet/write/mod.rs	51.28% <100.00%> (ø)
src/io/parquet/write/record_batch.rs	91.30% <100.00%> (ø)
tests/it/io/parquet/mod.rs	94.93% <100.00%> (ø)

---

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 8e96ec4...b7e2352. Read the comment docs.