Isaac Virshup

Can you provide us with more info about your reticulate environment? Like the version of `anndata`?

You could read through the sparse matrix and write it to the csv line by line, but if you can't hold the dense dataframe in memory it may be difficult...

> I can provide a Google Colab example soon to reproduce. That would be very useful, thanks!

Thanks! I can confirm the bug. Any chance you know what version of anndata wrote this file?

As a workaround for right now, if you write what you've read in with `anndata==0.7.8` the new release will be able to read that.

> Thanks ivirshup, so I can re-write the anndata and then read it correctly? Not sure how I would do it through kaggle, but I downloaded the dataset and can...

I think I might be okay with continuing to support this. I had explicitly removed support for files like this in 0.8.0 because I thought I would be the only...

Hey, this is expected. What you're looking for would be forward compatibility. Sometime we update the format of an AnnData objects stored on disk. We can't really make older versions...

Thanks for all the information! A number of the issues you raise are actually topics we're trying to address right now (and this release provides some solutions for), but it's...

Good resources to use here include: * [Benchmarking chapter from Jake VanderPlas' Python Data Science HandBook](https://jakevdp.github.io/PythonDataScienceHandbook/01.07-timing-and-profiling.html) * [loompy docs](http://linnarssonlab.org/loompy/index.html) * [Seurat docs](https://satijalab.org/seurat/index.html)