chris-b1

Thanks. Yep, in particular for me it is a messy data / evolving schema situation. E.g., data from Jan-Feb has columns `[:a, :b, :c]`, data from Mar has `[:a, :b,...

Got it, yeah not sure how hard this should be special cased. For what it's worth below is an isless comparer that got me to what I needed (sort with...

The actual usecase is missings last, because I'm using a descending sort. This would be pretty typical for reporting-type tasks - e.g., "show sales by region, ranked from largest to...

Just leaving an example here - this API is obviously not ideal, but this works for setting the numeric format for a particular type ```julia # at top level const...

gentle ping @bicycle1885 - any changes you'd like to see here?

As a workaround, you can downgrade to the last 1.x release of `xlrd` ```julia using Conda Conda.add("xlrd==1.2.0") ``` This could probably be pinned here - https://github.com/queryverse/ExcelReaders.jl/blob/master/src/ExcelReaders.jl#L12

I think something like https://github.com/JuliaLang/julia/issues/38162 would be necessary to fix this. It proposes "scoped" exports, so that names like `Router` could marked as API without being brought in as names...

Personally, I'd let the type conversion on mutation go away completely, I don't see any upside to that, especially with a first class NA value that doesn't force casts. More...

+1, although I think the opposite approach may also be worth consideration. What if instead of a being a special property of a `DataFrame`, an "Index" is just defined by...

No idea if this is a representative (probably not even particularly good code) but I was curious so here's a reduction microbenchmark. This is MSVC, with a only a 64...