Fixed error in reading nested parquet structs

[jorgecarleitao]

Closes #1014 - thanks @ahmedriza for the report!

[codecov]

Codecov Report

Merging #1015 (fb6765c) into main (bbe7209) will increase coverage by 0.00%.
The diff coverage is 73.40%.

@@           Coverage Diff           @@
##             main    #1015   +/-   ##
=======================================
  Coverage   71.68%   71.69%           
=======================================
  Files         359      359           
  Lines       19860    19880   +20     
=======================================
+ Hits        14237    14252   +15     
- Misses       5623     5628    +5     

Impacted Files	Coverage Δ
src/io/parquet/read/deserialize/binary/mod.rs	100.00% <ø> (ø)
src/io/parquet/read/deserialize/boolean/mod.rs	100.00% <ø> (ø)
src/io/parquet/read/deserialize/primitive/mod.rs	100.00% <ø> (ø)
src/io/parquet/read/deserialize/mod.rs	47.73% <57.69%> (+1.29%)	⬆️
src/io/parquet/read/deserialize/nested_utils.rs	74.47% <91.30%> (+0.25%)	⬆️
src/io/parquet/read/deserialize/struct_.rs	87.50% <93.75%> (+4.16%)	⬆️
src/io/parquet/read/deserialize/binary/nested.rs	61.25% <100.00%> (ø)
src/io/parquet/read/deserialize/boolean/nested.rs	62.74% <100.00%> (ø)
...rc/io/parquet/read/deserialize/primitive/nested.rs	73.61% <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 bbe7209...fb6765c. Read the comment docs.

[ahmedriza]

Thanks a lot @jorgecarleitao, I think this is still in progress. The reading looks good now. Getting an assertion failure during the write:

Writing /tmp/two_level_nested_verify.parquet
thread 'main' panicked at 'assertion failed: `(left == right)`
  left: `3`,
 right: `1`', /home/ahmed/.cargo/git/checkouts/arrow2-8a2ad61d97265680/f10a626/src/io/parquet/write/pages.rs:210:5

[jorgecarleitao]

That is expected - I need to improve the error message but the issue is that we pass vec![Encoding::Plain] to a field that has multiple columns - we require one encoding per parquet column.