The Kite Data module has the following known limitations as well as the intended course of action, if available. Users are encouraged to participate in the discussion of any feature addition or improvements by way of the mailing lists or the indicated JIRA, if it exists.
- It’s not currently possible to rename datasets.
- There’s no inter-process coordination or locking when operating on datasets. As a result, races can occur if two processes each try and create the same dataset.
- File names generated by the FileSystem* implementation writers are not guaranteed to be truly unique.
- All datasets are currently compressed using Snappy. It is not possible to disable compression or use a different codec.