So I take it you have zero experience with NoSQL databases then.
HiveQL is far more powerful than SQL and MongoDB QL is pretty powerful for querying document data structures. The others do 95% of what most people are doing in SQL. And the overwhelming majority of NoSQL databases support data types.
And there is no issue with the data integrity of NoSQL databases. Do you really think the hundreds of top companies would use them if there was e.g. Apple, Google, Twitter, Netflix etc.
Step 1: remove data integrity.
Step 2: remove powerful query language.
Step 3: only store strings.
=> No-SQL
It's interesting that the technology is named after a missing feature.