My pref is Postgres for its "good enough" text search and great JSON support. Most users nowadays expect a decent search functionality. And in PG it is pretty easy to add a ts_vector column for search purposes. I also find that JSON support alleviates the most common reason people want to use an ORM: a hierarchy of objects. Storing an object graph into tables is a real pain. Rather than using an ORM to numb that pain, I avoid it by storing those as JSON. PG supports indexes on jsonb fields as well. Not everything is best suited as JSON, but you can opt into using it where it makes sense. Otherwise PG is a pretty well-proven and robust SQL DB.
My pref is Postgres for its "good enough" text search and great JSON support. Most users nowadays expect a decent search functionality. And in PG it is pretty easy to add a
ts_vectorcolumn for search purposes. I also find that JSON support alleviates the most common reason people want to use an ORM: a hierarchy of objects. Storing an object graph into tables is a real pain. Rather than using an ORM to numb that pain, I avoid it by storing those as JSON. PG supports indexes onjsonbfields as well. Not everything is best suited as JSON, but you can opt into using it where it makes sense. Otherwise PG is a pretty well-proven and robust SQL DB.Hey! All your description sounds like CrateDB!!
But for me, both have them pros and cons.