ARROW-2554: [Python] fix timestamp unit detection from python lists

[crepererum]

No description provided.

[codecov-io]

Codecov Report

Merging #2118 into master will decrease coverage by 0.01%.
The diff coverage is 96%.

@@            Coverage Diff             @@
##           master    #2118      +/-   ##
==========================================
- Coverage   86.39%   86.38%   -0.02%     
==========================================
  Files         242      230      -12     
  Lines       41481    40611     -870     
==========================================
- Hits        35838    35080     -758     
+ Misses       5643     5531     -112

Impacted Files	Coverage Δ
python/pyarrow/tests/test_array.py	100% <100%> (ø)	⬆️
cpp/src/arrow/python/builtin_convert.cc	91.86% <94.44%> (+0.05%)	⬆️
rust/src/memory.rs
rust/src/builder.rs
rust/src/list.rs
rust/src/error.rs
rust/src/record_batch.rs
rust/src/array.rs
rust/src/memory_pool.rs
rust/src/buffer.rs
... and 4 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 6df28d3...850765e. Read the comment docs.

[xhochy]

LGTM except the cast.

[xhochy]

Use checked_cast<TimestampType&>(*type) here to avoid increasing the reference counter. (checked_cast is an Arrow-defined cast that evaluates to dynamic_cast in debug builds, otherwise it's a static_cast).

[wesm]

If you do that, make sure to use const auto& otherwise I believe the copy constructor will still be invoked

[wesm]

If the scalar type is some other unit (or unitless? can't remember how we're handling that at the moment) then you should raise in an else branch

[crepererum]

@xhochy ready :)

[xhochy]

+1, LGTM

Thanks!