Eric Hanson

I'm interested in what I think is the same problem in [Convex.jl](https://github.com/jump-dev/Convex.jl). I have been exploring rewriting some of the internals (ref https://github.com/jump-dev/Convex.jl/pull/393), and have come to the problem of...

Yep it was! Thanks, I might try that for Convex.jl too then.

The proper fix for this I think is to respect the difference between e.g. a 1x1 matrix and an scalar, but Convex.jl still has the MATLABism of conflating those all...

I believe https://github.com/jump-dev/Convex.jl/pull/504#discussion_r1434608140

I wonder if there's some way to use same redirected IO streams can be used for each `capture` invocation within a block, rather than separate streams each time[^1]. That would...

> This is almost assuredly a breaking change because deserialization behavior will change when TimeZones isn't loaded either directly or through some other package Regardless of breakingness, I'm not sure...

That sounds like a bug that the test is catching rather than just a flaky test!

I think the other answer is because (if I understand correctly) it's a DataFrames.jl feature, not an Arrow.jl feature, it's just documented here because it's a common ask

Maybe `collect` is what's desired here, for materializing a column into a `Vector`? Naively `collect` should be "iterate this collection into an `Array`", though I haven't tried it in this...

Oops, misclick