As for why, the rationale is that tabular data often needs more than the things that the article decries (e.g. efficient parsing / processing of partial records, sparse values, richly typed values, flexibility of schema evolution (i.e. not having to assume specific columns upfront)...
https://amzn.github.io/ion-docs/guides/cookbook.html#convert...
As for why, the rationale is that tabular data often needs more than the things that the article decries (e.g. efficient parsing / processing of partial records, sparse values, richly typed values, flexibility of schema evolution (i.e. not having to assume specific columns upfront)...