I have a loosely structured data(json) incoming from a source landing intno Snowflake stage. The structure of json(# of KV pairs) are not consistent across files i.e. one file can come up with 5 kv and another one can come up with 10.

How to handle such data in Snowflake if I plan to create wide table structure where each row represents data from a Json payload. If it was in parquet, it would be a sparse table with column values being null values when they are missing from the source.

How does Snowflake handle such inputs. Does it create sparse tables as well, and can you also advice on the schema evolution. Any gotchas as # of columns are updated with each payload.