Remove len parameter from ScalarFnArray::try_new

[connortsui20]

Summary

This didn't need to take length because it already gets it from the array children.

API Changes

Removes len parameter.

Testing

N/A

[codspeed-hq]

Merging this PR will improve performance by 36.9%

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

⚡ 3 improved benchmarks
✅ 1523 untouched benchmarks
⏩ 10 skipped benchmarks¹

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	Simulation	chunked_bool_canonical_into[(1000, 10)]	35.5 µs	20.5 µs	+73.18%
⚡	Simulation	chunked_varbinview_canonical_into[(1000, 10)]	198.1 µs	161.5 µs	+22.63%
⚡	Simulation	chunked_varbinview_into_canonical[(1000, 10)]	213.5 µs	176.8 µs	+20.81%

Tip

Curious why this is faster? Comment @codspeedbot explain why this is faster on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing ct/remove-len-scalar-fn-array (ea3ad14) with develop (8475910)}

Footnotes

1. 10 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]

@claude fix the merge conflicts

[github-actions]

Claude finished @connortsui20's task in 3m 54s —— View job

---

Merge conflict analysis

• Fetch latest develop and identify conflicts
• Determine the correct resolution
• Resolve conflicts and push — blocked by environment restrictions (see below)

---

What's conflicting

develop has advanced since this branch was created. Its current HEAD is "Remove all TurboQuant stuff" (#8377), which deletes the entire vortex-turboquant crate and all TurboQuant/tensor code. This PR modifies many of those now-deleted files, producing modify/delete conflicts.

Of the 25 files in this PR, 15 no longer exist on develop and should simply be dropped:

Deleted on develop (drop these)
vortex-turboquant/src/scalar_fns/decode.rs, encode.rs
vortex-turboquant/src/vector/normalize.rs
vortex-tensor/src/encodings/turboquant/compress.rs
vortex-tensor/src/encodings/turboquant/tests/{compute,nullable,structural}.rs
vortex-tensor/src/scalar_fns/{cosine_similarity,inner_product,l2_denorm,l2_norm}.rs
vortex-tensor/src/scalar_fns/sorf_transform/{mod,tests}.rs
vortex-tensor/src/{utils,vector_search}.rs

The remaining 10 files carry the actual API change (removing len from ScalarFnArray::try_new) and still exist on develop, so they should be kept:

• vortex-array/src/arrays/scalar_fn/{array,plugin,rules}.rs
• vortex-array/src/arrays/scalar_fn/vtable/operations.rs
• vortex-array/src/arrays/chunked/compute/rules.rs
• vortex-array/src/arrays/dict/compute/rules.rs
• vortex-array/src/arrays/fixed_size_list/tests/nested.rs
• vortex-array/src/expression.rs
• encodings/datetime-parts/src/compute/rules.rs
• encodings/runend/src/rules.rs

Why I couldn't push the fix

This review environment denies the git operations required to build the merge commit:

• The checkout is a shallow clone, so there's no reachable merge base between this branch and origin/develop.
• git fetch --deepen / --unshallow (network) and git merge are both blocked / require approval that isn't grantable here — even with the sandbox disabled.

So I can't create or push the merge commit from this job. To enable it, add git fetch and git merge to --allowedTools (and allow network), or resolve locally:

Recommended resolution (≈1 min locally)

git switch ct/remove-len-scalar-fn-array
git fetch origin develop
git merge origin/develop
# All conflicts are modify/delete on removed TurboQuant/tensor files — accept the deletions:
git rm vortex-turboquant/src/scalar_fns/decode.rs \
       vortex-turboquant/src/scalar_fns/encode.rs \
       vortex-turboquant/src/vector/normalize.rs \
       vortex-tensor/src/encodings/turboquant/compress.rs \
       vortex-tensor/src/encodings/turboquant/tests/compute.rs \
       vortex-tensor/src/encodings/turboquant/tests/nullable.rs \
       vortex-tensor/src/encodings/turboquant/tests/structural.rs \
       vortex-tensor/src/scalar_fns/cosine_similarity.rs \
       vortex-tensor/src/scalar_fns/inner_product.rs \
       vortex-tensor/src/scalar_fns/l2_denorm.rs \
       vortex-tensor/src/scalar_fns/l2_norm.rs \
       vortex-tensor/src/scalar_fns/sorf_transform/mod.rs \
       vortex-tensor/src/scalar_fns/sorf_transform/tests.rs \
       vortex-tensor/src/utils.rs \
       vortex-tensor/src/vector_search.rs
# The 10 vortex-array / encodings files carrying the `len` removal should merge cleanly.
git commit          # completes the merge
git push

Note

Given that #8377 removed the whole TurboQuant subsystem, those 15 files in this PR were only call-site updates for code that no longer exists — dropping them loses nothing. The meaningful change (the try_new signature) lives entirely in the 10 vortex-array/encodings files. You may even want to retitle/trim the PR after the merge, since over half its diff disappears.
· branch ct/remove-len-scalar-fn-array

[joseph-isaacs]

What is the scalar fn doesn't have children?

pub fn try_new(scalar_fn: ScalarFnRef, children: Vec<ArrayRef>) if the vec.len is empty?

[robert3005]

There’s still a method that lets you provide len. I think it’s nicer since the other cases are very few compared to cases with non empty children

[joseph-isaacs]

No cause now if the children is empty this becomes a bail or panic! For no reason

[robert3005]

This is the same thing as constructing chunkedarray from children and dtype or just children, having empty children panics.