ARROW-2589: [Python] Workaround regression in Pandas 0.23.0

[pitrou]

There is a regression (*) in Pandas 0.23.0 that breaks test_parquet.py.
Pandas does not have an actual "str" dtype anyway, so pass "object" instead.

(*) pandas-dev/pandas#21083

[pitrou]

Another possible fix is to replace the "str" dtype with "object", since "str" seemed to be a synonym of "object" before Pandas 0.23.0.

@TomAugspurger what do you think?

[TomAugspurger]

For test-code like in https://travis-ci.org/apache/arrow/jobs/379652352#L3156, that should be a fine workaround for now.

In the future when pandas has a dedicated string type, object and str will no longer be synonymous.

pandas-dev/pandas#8640 for tracking that (it may happen soon).

[codecov-io]

Codecov Report

Merging #2051 into master will increase coverage by 0.02%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #2051      +/-   ##
==========================================
+ Coverage   87.44%   87.47%   +0.02%     
==========================================
  Files         189      178      -11     
  Lines       29374    28601     -773     
==========================================
- Hits        25686    25018     -668     
+ Misses       3688     3583     -105

Impacted Files	Coverage Δ
cpp/src/arrow/util/thread-pool-test.cc	98.87% <0%> (-0.57%)	⬇️
rust/src/builder.rs
rust/src/list.rs
rust/src/memory.rs
rust/src/error.rs
rust/src/buffer.rs
rust/src/memory_pool.rs
rust/src/datatypes.rs
rust/src/bitmap.rs
rust/src/array.rs
... and 2 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 6ca246a...b581ef3. Read the comment docs.

[xhochy]

+1, LGTM