PARQUET-1494: [C++] Recognize statistics built with UNSIGNED sort order by parquet-mr 1.10.0 onwards

[zilder]

Fixes the issue with min-max statistics built by parquet-mr 1.10+:
https://issues.apache.org/jira/browse/PARQUET-1494

[majetideepak]

Should be parquet-mr

[zilder]

sure, sorry

[majetideepak]

Edit: Can you add the file generated by parquet-mr to the data folder in the parquet-testing submodule https://github.com/apache/arrow/tree/master/cpp/submodules and add a test as well?

[zilder]

Added a test case.

[zilder]

Btw, thanks to @sumerman for helping to investigate the issue and @alkagin for providing the sample parquet file for the test case.

[majetideepak]

@zilder can you amend your last commit and push again? I committed the binary file. Travis CI should pick that up now.

[majetideepak]

@zilder looks like you need to run clang-format

/home/travis/build/apache/arrow/cpp/src/parquet/statistics-test.cc had clang-format style issues
--- /home/travis/build/apache/arrow/cpp/src/parquet/statistics-test.cc
+++ /home/travis/build/apache/arrow/cpp/src/parquet/statistics-test.cc (after clang format)
@@ -777,7 +777,8 @@
 TEST(TestStatisticsMinMax, Unsigned) {
   std::string dir_string(test::get_data_dir());
   std::stringstream ss;
-  ss << dir_string << "/" << "binary.parquet";
+  ss << dir_string << "/"
+     << "binary.parquet";

[zilder]

just did and pushed new version

[majetideepak]

+1 LGTM. I will merge after appveyor finishes. Thanks for this contribution!

[wesm]

@majetideepak has someone flipped your commit bit for arrow? Apologies if not, Jacques may have to do it

[majetideepak]

@wesm I do not have the commit bit for Arrow. I will send an email to Jacques. Can you take care of merging this? Thanks.