ANN: LightQuery tutorial

Ok, I’ll test this today, thanks.

I’m just curious, are there ever situations where “a join” wouldn’t be that 5-row table…? Like, what is the advantage of having to sort, group, and then mix? I can understand the modularity and how powerful that can be, but a join is a join. So the convenience of a normal join function couldn’t possibly be underestimated. But that might be just me.

I’ll just add another thing I noticed: you Group only the first source and you order none of the sources. I’m sure this makes sense, it just seems a bit difficult to penetrate this API from my (I guess “a beginner”) point of view. Consider that we have about 10 posts in this topic just about how to do a simple inner join – this probably says more about how slow I am than anything else. It also seems like it takes about 6×2+1 LOC to accomplish that join. This probably depends on the fact that you can streamline and perfect any action depending on the state of the input, which makes LightQuery the ideal building block for all the query frontend packages.

But all of this is a good indication that we need either:

  1. a convenience function to accomplish (albeit not-as-fast-and-streamlined-as-possible) joins (and other DB actions) that never fail,
  2. or a MetaLightQuery package for day-to-day DB actions (like join).

Again, I hope you don’t see this in the wrong light (heh), I love the modularity and lightness, but if normal users are to use this package, my opinion is such that this needs some convenience functions.

1 Like

I’d love for a MetaLightQuery. The point of the package is really to be low-level but flexible tools for working with tabular data. The reason I didn’t order the sources is because in your example they were already in order.

this probably says more about how slow I am than anything else

I think it probably just says that I’m a terrible writer. I’d love any help/advice for how to write better documentation.