Add defensive checks for Parquet statistics handling and improve library path resolution

[a-hirota]

Summary

Adds defensive checks for Parquet statistics handling to prevent crashes when ABI
incompatibilities exist between compile-time and runtime Arrow/Parquet library
versions.

Background

Issue #940 reported segmentation faults when reading Parquet files with statistics
created by PyArrow 19.0.1. Investigation revealed that the crashes occurred due to
ABI mismatches between different versions of Arrow/Parquet libraries loaded at
compile-time vs runtime.

Changes

arrow_meta.cpp

• Add null pointer checks for statistics objects before accessing methods
• Wrap physical_type() and statistics method calls in try-catch blocks
• Gracefully handle exceptions from virtual method calls on corrupted vtables

Makefile

• Add automatic rpath configuration for Arrow/Parquet libraries using pkgconf
• Ensures runtime library path matches compile-time library selection

Testing

• ✅ Successfully reads PyArrow 19.0.1 created Parquet files with statistics
• ✅ Handles mixed data types (int64, string, double) without crashes
• ✅ Graceful degradation when statistics cannot be accessed

Impact

While the root cause (library version conflicts) should be resolved at the
environment level, these defensive changes provide:

• Crash prevention in mixed-version environments
• Better error handling for statistics processing
• Improved library path resolution

Fixes #940

[Copilot]

Pull Request Overview

This PR adds defensive error handling to prevent segmentation faults when reading Parquet files with statistics created by different Arrow/Parquet library versions, addressing issue #940.

• Added null pointer checks and exception handling for Parquet statistics method calls
• Wrapped statistics access in try-catch blocks to handle ABI incompatibilities gracefully
• Enhanced Makefile to automatically configure runtime library paths using pkgconf

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
src/arrow_meta.cpp	Added defensive checks for statistics access with try-catch blocks and null pointer validation
src/Makefile	Added automatic rpath configuration for Arrow/Parquet libraries to ensure runtime path matches compile-time selection

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

The code still calls stats->physical_type() without additional protection after the null check. If the object is corrupted due to ABI mismatch, this call itself could segfault before the exception handlers can catch it. Consider adding a more defensive approach or additional validation before making any method calls on the stats object.

[Copilot]

Similar to the issue in __readParquetMinMaxStats, the method calls stats->HasNullCount() and stats->HasMinMax() could still segfault due to corrupted vtables before the exception handlers can catch them. These calls should be moved inside the try block or additional validation should be added.

[a-hirota]

I've addressed the Copilot suggestions regarding defensive programming for Parquet
statistics handling:

Changes Made

1. Enhanced Exception Handling Granularity

• Separated null count and min/max statistics into individual try-catch blocks
• This allows partial recovery: if one type of statistics fails, the other can still
  be processed
• Each failure type now has specific debug logging for better troubleshooting

2. Improved Fault Isolation

// Before: Single try-catch for all statistics methods
try {
    if (stats->HasNullCount()) field->null_count = stats->null_count();
    if (stats->HasMinMax()) __readParquetMinMaxStats(field, stats);
} catch (...) { /* handle all failures together */ }

// After: Individual protection for each statistics type
try {
    if (stats->HasNullCount()) field->null_count = stats->null_count();
} catch (...) { /* handle null count failures specifically */ }

try {
    if (stats->HasMinMax()) __readParquetMinMaxStats(field, stats);
} catch (...) { /* handle min/max failures specifically */ }

3. Benefits of This Approach
- Partial recovery: If HasNullCount() fails due to ABI mismatch, min/max statistics
can still be processed
- Better debugging: Specific error messages help identify which statistics method is
affected
- Graceful degradation: Maximum data extraction even in mixed-version environments

This addresses the Copilot concern about insufficient protection while maintaining
performance and providing better error isolation for debugging ABI compatibility
issues.