Support NA values with boolean column (fix #100, fix #102, fix #103)

[laughingman7743]

No description provided.

[codecov]

Codecov Report

Merging #104 into master will increase coverage by 0.39%.
The diff coverage is 90.47%.

@@            Coverage Diff             @@
##           master     #104      +/-   ##
==========================================
+ Coverage   91.99%   92.39%   +0.39%     
==========================================
  Files          14       14              
  Lines        1062     1104      +42     
==========================================
+ Hits          977     1020      +43     
+ Misses         85       84       -1

Impacted Files	Coverage Δ
pyathena/result_set.py	92.59% <ø> (-0.22%)	⬇️
pyathena/connection.py	88.6% <100%> (+1.1%)	⬆️
pyathena/util.py	100% <100%> (+11.53%)	⬆️
pyathena/converter.py	86.95% <88.63%> (+6.62%)	⬆️
pyathena/formatter.py	94.93% <88.88%> (-2.04%)	⬇️

---

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 3eef7d2...03805e5. Read the comment docs.

[tanguycdls]

👍 Nice tests ! I understand that since you dont provide a dtype for the boolean column when calling read_csv. Pandas cast it either to boolean if there are no missing values or object if there are some ?

That would explain why it's a None and not a na.

Thanks a lot for the prompt PR :)

[laughingman7743]

I wrote how to customize dtypes and converters in README.

[laughingman7743]

PyAthena/pyathena/converter.py

Line 145 in 6276338

class DefaultPandasTypeConverter(Converter):

By default this class is used.