I enjoy using mongo / other document stores, but this article is lame. The contrast is between a single document storing an order, with sub-documents containing details, versus a table per sub-doc type. They claim you have to denormalize data in the relational version, which isn't true, and use that claim to assert that updates are complicated.
Guess what? Denormalizing data can make things complicated in mongo-land, too. This isn't just a relational db problem.
Similarly, the order lines don’t contain product names, but rather foreign keys which point to a product table
Well, the product name could be the foreign key. Or it could not be. If you have 2 different products with the same name (or 1 person with 2 different names) then you have some complexity. Doesn't matter what kind of database you're using.
I don't believe it's a straw man, and you haven't refuted the posters view other than to say you don't believe the historical preservation is necessary without elaboration.
I'd also point out that some relational luminaries advocate a similar perspective, such as Pat Helland.
Comments
I enjoy using mongo / other document stores, but this article is lame. The contrast is between a single document storing an order, with sub-documents containing details, versus a table per sub-doc type. They claim you have to denormalize data in the relational version, which isn't true, and use that claim to assert that updates are complicated.
Guess what? Denormalizing data can make things complicated in mongo-land, too. This isn't just a relational db problem.
Yeah, total straw man. For e.g.:
Similarly, the order lines don’t contain product names, but rather foreign keys which point to a product table
Well, the product name could be the foreign key. Or it could not be. If you have 2 different products with the same name (or 1 person with 2 different names) then you have some complexity. Doesn't matter what kind of database you're using.
I don't believe it's a straw man, and you haven't refuted the posters view other than to say you don't believe the historical preservation is necessary without elaboration.
I'd also point out that some relational luminaries advocate a similar perspective, such as Pat Helland.