ARROW-3728: [Python] Ignore differences in schema custom metadata when writing table to ParquetWriter

[kszucs]

Merging tables with schemas identical field-wise but different in metadata fails.

[wesm]

Can you explain the issue in the PR description? Thank you :)

[kszucs]

Do We want to provide an API for merging parquet files?

[codecov-io]

Codecov Report

Merging #3029 into master will increase coverage by 1.04%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #3029      +/-   ##
==========================================
+ Coverage   86.99%   88.03%   +1.04%     
==========================================
  Files         494      425      -69     
  Lines       70410    64819    -5591     
==========================================
- Hits        61253    57066    -4187     
+ Misses       9061     7753    -1308     
+ Partials       96        0      -96

Impacted Files	Coverage Δ
python/pyarrow/parquet.py	93.89% <100%> (+0.19%)	⬆️
python/pyarrow/tests/test_parquet.py	97.39% <100%> (+0.02%)	⬆️
cpp/src/arrow/util/thread-pool-test.cc	98.91% <0%> (-0.55%)	⬇️
rust/src/record_batch.rs
go/arrow/array/table.go
rust/src/array.rs
go/arrow/math/uint64_amd64.go
go/arrow/internal/testing/tools/bool.go
go/arrow/array/bufferbuilder.go
go/arrow/internal/bitutil/bitutil.go
... and 64 more

---

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 7281731...2e12f24. Read the comment docs.

[wesm]

Could you clarify what you mean by "merging parquet files"? Like reading multiple files and writing out a new file? It might be useful, feel free to open a JIRA issue

[wesm]

+1, thanks! I opened https://issues.apache.org/jira/browse/ARROW-3876 to think more about schema normalization on write in general. We need to be careful about preserving metadata without being too much of a nuisance to the user