Use kernel utility for parsing timestamps in csv reader.

[novemberkilo]

Which issue does this PR close?

This is related to a datafusion issue -- apache/datafusion#958

Rationale for this change

apache/datafusion#958 (comment)

What changes are included in this PR?

So as to support more timestamp formats, this PR changes the implementation of parse for Microsecond and Nanosecond type Timestamps to use kernels::cast_utils::string_to_timestamp_nanos

Are there any user-facing changes?

I've checked that a user can now parse timestamps in CSV files as described in the referenced datafusion issue. Specifically, you can now do:

fn nyctaxi_schema() -> Schema {
    Schema::new(vec![
        Field::new("VendorID", DataType::Utf8, true),
        Field::new("pickup_datetime", DataType::Timestamp(TimeUnit::Microsecond, None), true),
        Field::new("dropoff_datetime", DataType::Timestamp(TimeUnit::Microsecond, None), true),
        ...

// @alamb

[codecov-commenter]

Codecov Report

Merging #832 (4aad3dc) into master (7472e5e) will increase coverage by 0.03%.
The diff coverage is 98.07%.

@@            Coverage Diff             @@
##           master     #832      +/-   ##
==========================================
+ Coverage   82.60%   82.64%   +0.03%     
==========================================
  Files         168      168              
  Lines       48040    48088      +48     
==========================================
+ Hits        39685    39742      +57     
+ Misses       8355     8346       -9     

Impacted Files	Coverage Δ
arrow/src/csv/reader.rs	91.77% <98.07%> (+1.86%)	⬆️

---

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 7472e5e...4aad3dc. Read the comment docs.

[alamb]

Looks good to me -- thank you @novemberkilo -- this change will ensure that the timestamp format supported by the csv parser are the same as the timestamp formats supported by the cast kernel (utf8 -> timestamp) 👍

[alamb]

I think this PR will pass CI checks once it is rebased against master to pick up #839

[alamb]

I don't think this stringify is needed here: it converts naive_datetime into the string -- so this message always looks like:

thread 'main' panicked at 'Unexpected failure converting naive_datetime to local datetime', src/main.rs:2:3

Which you can see at the playground: https://play.rust-lang.org/?version=stable&mode=debug&edition=2018&gist=9f6fbabade9152946ea5a84db486c2de

Since it is in test code, this is likely not a critical problem but I noticed it so figured I would point it out

[novemberkilo]

Thanks for pointing that out Andrew. Will fix.

[alamb]

Thanks again @novemberkilo !