Data loading may have side effects due to naming conflicts

[gbrgr]

In the functions load_csv and load_json, we use an auxiliary relation named data to populate the configuration relation for data loading. However, if one's workspace already contains data within the relation data, these operations error out and sources of error are hard to find.

Notice that we cannot yet create an anonymous relation for config in the call load_csv[config], as this is not supported yet for FFI calls. A simple solution would be to generate a random relation name which has little probability for conflicts.

[billscheidel-rai]

Note: This issue has been migrated to https://relationalai.atlassian.net/browse/RAI-5169.

This link is only accessible to employees of RelationalAI.