Re:Many small queries are efficient in SQLite
by dragonwriter on Tuesday February 16, 2021 @06:01AM
(#26152820)
> My understanding is that SQLite doesn't impose any format either?
That's true that you can stuff any kind of string/blob data into any column of any table, so, yes, you still have to determine the data schema with sqlite much as you do with JSON, XML, or even CSV. I mean, I could have a CSV where each element is a base64-encoded ZIP containing sqlite database files that are each a single table with a single column of JSON files, each of which contains a JSON array of strings with XML documents in them.
But that's usually not something people would mean if they said their app was using CSV as it's data storage format, nor is the version stripping out CSV on the top what people would mean if they say they are using SQLite.
With ZIP, you have to decide the format(s) for the file(s) in the ZIP, their hierarchical structure, and, if the files aren't themselves the atomic data elements, the schema applicable to each file.
Furthermore, in discussion of performance characteristics and other aspects of suitability, ZIP adds overhead, but you still also need to consider the access properties of the contained files.
That's true that you can stuff any kind of string/blob data into any column of any table, so, yes, you still have to determine the data schema with sqlite much as you do with JSON, XML, or even CSV. I mean, I could have a CSV where each element is a base64-encoded ZIP containing sqlite database files that are each a single table with a single column of JSON files, each of which contains a JSON array of strings with XML documents in them.
But that's usually not something people would mean if they said their app was using CSV as it's data storage format, nor is the version stripping out CSV on the top what people would mean if they say they are using SQLite.
With ZIP, you have to decide the format(s) for the file(s) in the ZIP, their hierarchical structure, and, if the files aren't themselves the atomic data elements, the schema applicable to each file.
Furthermore, in discussion of performance characteristics and other aspects of suitability, ZIP adds overhead, but you still also need to consider the access properties of the contained files.