table.find should allow selecting specific columns

[will123195]

Like mongo projections

db.authors.find({
  columns: ['name', 'city'], // or maybe { name: 1, city: true }
  where: { name: 'Will' }
})

[rneilson]

It seems like .hydrate() would be the place to implement this, since .find() uses it under the hood (if I'm reading things right)?

[will123195]

.find() and .hydrate() both use .get() under the hood. So we could add a columns option like this:

movies.get(123, { 
  columns: { title: 'myTitle' }
})
.then(movie => {
  // { id: 123, myTitle: 'Office Space' }
})

I guess hydrate could be supported like this:

movies.get(123, { 
  hydrate: { director: { columns: { name: true } } }
})
.then(movie => {
  // { id: 123, title: 'Office Space', year: 1999, director: { id: 1, name: 'Mike Judge' } }
})

I'm thinking the primary key should be automatically included. Otherwise movie.director.save() will not work.

[rneilson]

I think I know how I'd put this in, but there are the questions of instantiation and caching:

• If a subset of fields are retrieved, will that interfere with instantiating the Row object? (Or just return a plain object instead?)
• Should updating the cache be skipped? (I'm presuming so, since I don't think it'd make sense. Unless there's room for some kind of qualifier in the cache to specify the field subset.)

[will123195]

• I think partial Row instantiation and partialRow.save() should work fine.
• I agree let's not cache partial rows