Added support to read dict-encoded required primitive types from parquet

[Dandandan]

Closes #400

[codecov]

Codecov Report

Merging #402 (a9d980b) into main (227ab3b) will increase coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main     #402      +/-   ##
==========================================
+ Coverage   80.90%   80.91%   +0.01%     
==========================================
  Files         347      347              
  Lines       22086    22098      +12     
==========================================
+ Hits        17869    17881      +12     
  Misses       4217     4217              

Impacted Files	Coverage Δ
src/io/parquet/read/primitive/basic.rs	85.00% <100.00%> (+2.14%)	⬆️
tests/it/io/parquet/read.rs	98.50% <100.00%> (+0.04%)	⬆️

---

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 227ab3b...a9d980b. Read the comment docs.

[jorgecarleitao]

Awesome!

I think we are just missing a small test here. I have been using pyarrow for this. Specifically,

1. write a parquet (arrow-parquet-integration-testing) with dict-encoded values
2. have the expected values in fn pyarrow_required in tests/it/io/parquet/mod.rs
3. perform a read test for the corresponding column and file at tests/it/io/parquet/read.rs

I think 1. and 2. are done, we just need to add something like

#[test]
fn v2_int64_required_dict() -> Result<()> {
    test_pyarrow_integration(0, 2, "basic", true, true)
}

#[test]
fn v1_int64_required_dict() -> Result<()> {
    test_pyarrow_integration(0, 1, "basic", true, true)
}

to check that we can read parquet files written by pyarrow, in both versions.

[jorgecarleitao]

Released in v0.5.3.

[Dandandan]

Thanks!