Bulk take in ListView naive_rebuild

[palaska]

Replace per-list extend_from_array(slice(...)) with a single take in naive_rebuild.

The old approach canonicalized elements upfront, then called extend_from_array per list — each call triggered buffer cloning for VarBinView. The new approach collects element indices into a BufferMut<u64> and
does one bulk take.

Benchmarks (10k lists, overlapping views):

• Primitives: 2,006 µs → 79 µs (~25x)
• VarBinView: 3,972 µs → 74 µs (~52x)

[codspeed-hq]

Merging this PR will improve performance by ×8.5

⚡ 1 improved benchmark
✅ 1134 untouched benchmarks
🆕 6 new benchmarks
⏩ 1268 skipped benchmarks¹

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
🆕	Simulation	rebuild_varbinview[1000]	N/A	338.8 µs	N/A
🆕	Simulation	rebuild_primitive[10000]	N/A	2.5 ms	N/A
🆕	Simulation	rebuild_primitive[1000]	N/A	295.5 µs	N/A
🆕	Simulation	rebuild_primitive[100]	N/A	71.1 µs	N/A
🆕	Simulation	rebuild_varbinview[100]	N/A	102.2 µs	N/A
🆕	Simulation	rebuild_varbinview[10000]	N/A	4.4 ms	N/A
⚡	Simulation	rebuild_naive	2,718 µs	321 µs	×8.5

---

_{Comparing bp/listview-rebuild (dd49507) with develop (5de9826)}

Footnotes

1. 1268 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[connortsui20]

very nice! LGTM

[joseph-isaacs]

this builds a lazy take plan, do you want this?

[joseph-isaacs]

I think you want a execute here?

[joseph-isaacs]

I think you want a execute here?

[joseph-isaacs]

Its faster if you don't do the work 🙈

[connortsui20]

oh right all of our compute is lazy now haha

But I'll be surprised if this is not significantly better than the old naive version for primitive types

Edit: Actually are we sure we want to execute here? I don't think it is absolutely necessary to do so, especially if we want to filter/take after a rebuild has occurred (in which case we can optimize that plan)

[palaska]

Yeah my bad, didn't know take was lazy. It's still a lot faster than previous impl though

[connortsui20]

In the case that we read ListViewArray from disk and we want to export to datafusion (basically if we want to export as a ListArray to Arrow), I believe that not executing here might be more beneficial.

If we want to push down a filter/take into a ListView from Datafusion/arrow, then we definitely do not want to execute here since it would be better to combine the filter/take with this take here.

I also think that this is the more common (perhaps only?) path that this naive_rebuild exists in.

cc @joseph-isaacs

[palaska]

Yeah, not sure what the caller expects to get here. I think it makes sense to keep it lazy and leave canonicalization up to the caller, wdyt @joseph-isaacs ?