fix(parquet): drain lazy preloadedRepDefs_ from seekToPage

[JinyuanZhang617]

Background

For nested (non top level) columns Bolt only fully decodes the first decodeRepDefPageCount_ (default 10) pages of each column chunk in preloadRepDefs(). The remaining pages are kept as raw rep/def bytes inside preloadedRepDefs_ and decoded on demand by loadMoreRepDefs(). guhaiyan@ commit 27e789e added an ahead-by-one-page invariant at the exit of decodeRepDefs() so that the next consumer access can always read numLeavesInPage_[pageIndex_] safely.

That invariant only fires on the read path driven by RepeatedColumnReader::readRepeatedFor. When a sibling top-level filter pushdown causes the nested column leaf reader to take the skip path instead, control flows SelectiveColumnReader::skip -> PageReader::skip -> seekToPage -> prepareDataPageV1 -> setPageRowInfo and never touches decodeRepDefs. If the skip distance is large enough to cross the sampled boundary, setPageRowInfo does ++pageIndex_ followed by numLeavesInPage_[pageIndex_] and raises (10 vs. 10) Seeking past known repdefs for non top level column page 10.

Fix

Hoist the lazy-load loop into seekToPage page-walk:

• fired only when hasChunkRepDefs_ AND not isTopLevel_ AND maxRepeat_ greater than 0, so the top-level / no-repdef paths stay zero-cost
• uses pageIndex_ + 1 >= numLeavesInPage_.size() so we always have the per-page leaf count for the page we are about to enter, mirroring guhaiyan@ ahead-by-one-page invariant on the read path
• setPageRowInfo stays a pure lookup, keeping the original layering between rep/def producer (preloadRepDefs / loadMoreRepDefs) and consumer (setPageRowInfo / getLengthsAndNulls)
• loadMoreRepDefs is called at most once per page header, vs. an earlier attempt where it could be called per setPageRowInfo call

Regression test

lazyRepDefSkipPastSampledBoundary writes a single row group containing a sparse MAP STRING STRING column with dataPageSize=1024 so the chunk has many small pages, applies a sparse IN filter on a sibling BIGINT column and lowers decodeRepDefPageCount to 2 via RowReaderOptions to make the sampling boundary deterministic. Without this change the test reproduces the assertion failure; with the change the read completes successfully.

Change-Id: Iaec5fb28f07533762c1421cba0601096eb203fe4

What problem does this PR solve?

Issue Number: close #xxx

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

Describe your changes in detail.
For complex logic, explain the "Why" and "How".

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:
- Fixed a crash in `substr` when input is null.
- optimized `group by` performance by 20%.

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]

where is the fix? comments only ?