sqlhbenchmarks VS bqb

Another choice is a library like squirrel which helps with building SQL statements, adding parameters, and execution. Such libraries tend to be a bit memory hungry though. I have benchmarked some libraries against each other awhile back here https://github.com/nofeaturesonlybugs/sqlhbenchmarks If you look at the benchmarks involving squirrel you can see it tends to make many more allocations than the alternatives I tested.

There's a sibling package containing benchmarks @ https://github.com/nofeaturesonlybugs/sqlhbenchmarks

But using a query builder, something like squirrel or (plug) bqb, allows you to actually write SQL (or something close to it) when you need it but also handles the nasty string building bits. Though I agree that ORMs are not always bad, especially for small projects with well-defined scope.

Plug for bqb as a query builder, but there's also squirrel which works pretty well too.

(plug) bqb for very simple query building.

Squirrel is great! Let me also plug bqb.

Plug: BQB (basic query builder) is small, 100% test coverage, and in AwesomeGo. A great starter project.

I totally understand this. If you do go with a query builder, may I recommend bqb (shameless plug) as it allows you to remain closer to the SQL than some alternatives and doesn't do anything fancy. We stripped out 50% of our query logic with it in our org, which has enabled us to more readily tweak the SQL for performance.

There's also bqb. We use it in production at our company -- much better than raw SQL. If you couple it with something like scany then you get more of the ORM benefits without the complexity.

Another shameless plug: https://github.com/nullism/bqb pretty tiny query builder with 100% test coverage