In Dask Array, is there any difference between a block and a chunk?

Viewed 191

In Dask Array, is there any difference between a block and a chunk?

In the docs it seems like the two are used synonymously. E.g. here (at list item #3 they say block instead of chunk) or here.

1 Answers

Chunks are basically the same thing as blocks for arrays. But chunk in this context is more like an action: to cut the whole dataset into chunks/blocks.

This action is also called partitioning in some contexts, particularly for dataframes and bags (which also have chunks, but they are not normally called blocks).

Related