Add a Flush API to the Parquet writer to flush the buffered row group.

[luozenglin]

What problem does this PR solve?

Issue Number: close #212

Type of Change

• 🐛 Bug fix (non-breaking change which fixes an issue)
• ✨ New feature (non-breaking change which adds functionality)
• 🚀 Performance improvement (optimization)
• ⚠️ Breaking change (fix or feature that would cause existing functionality to change)
• 🔨 Refactoring (no logic changes)
• 🔧 Build/CI or Infrastructure changes
• 📝 Documentation only

Description

Add a Flush API to the Parquet writer to flush the buffered row group.

In scenarios where BufferedRowGroup is used, the current NewBufferedRowGroup API flushes the current row group and also creates a new BufferedRowGroup. In some cases, we only want to flush the current row group without creating a new BufferedRowGroup, to avoid writing an empty row group.

Performance Impact

• No Impact: This change does not affect the critical path (e.g., build system, doc, error handling).

• Positive Impact: I have run benchmarks.

  Click to view Benchmark Results

  Paste your google-benchmark or TPC-H results here.
  Before: 10.5s
  After:   8.2s  (+20%)
  

• Negative Impact: Explained below (e.g., trade-off for correctness).

Release Note

Please describe the changes in this PR

Release Note:

Release Note:
- Add a `Flush` API to the Parquet writer to flush the buffered row group.

Checklist (For Author)

• I have added/updated unit tests (ctest).
• I have verified the code with local build (Release/Debug).
• I have run clang-format / linters.
• (Optional) I have run Sanitizers (ASAN/TSAN) locally for complex C++ changes.
• No need to test or manual test.

Breaking Changes

• No

• Yes (Description: ...)

  Click to view Breaking Changes

  Breaking Changes:
  - Description of the breaking change.
  - Possible solutions or workarounds.
  - Any other relevant information.
  

[guhaiyan0221]

why a temp assignment needed ?
PARQUET_CATCH_NOT_OK(row_group_writer_->Close()); row_group_writer_ = nullptr; does not work ?

[luozenglin]

This ensures that row_group_writer_ is set to nullptr when an exception occurs. When row_group_writer_->Close() fails (e.g., due to an HDFS write error), the upper layer typically calls Close() to terminate writing this Parquet file. If we don’t reset row_group_writer_ to nullptr, the Close() method will invoke row_group_writer_->Close() again. At that point, the data inside row_group_writer_ may already be incomplete and it may throw again; calling row_group_writer_->Close() is no longer meaningful.

[guhaiyan0221]

Better to use Guard or try-catch(...) to deal with this case.

[guhaiyan0221]

Why is this change needed?

[luozenglin]

When enableFlushBasedOnBlockSize_ is enabled, the underlying implementation uses BufferedRowGroup, which only takes effect if the Flush API is called. This change ensures the correctness of the Writer::flush semantics and also enables adding tests.

[guhaiyan0221]

Better to use Guard or try-catch(...) to deal with this case.