Why are grams usually only expressed as milligrams, grams or kilograms? I'm a physics (and electronics and astronomy, etc.) enthusiast. As I learn and research topics, I notice that many SI units are often expressed using a variety of prefixes, such as in electronics where we use microvolts, millivolts, volts, kilovolts, and sometimes megavolts. In computer storage, the prefixes kilo, mega, giga, and tera are very familiar.
In physics, however, for large quantities of mass, I usually just see kilograms used with scientific notation:
$$2\times10^6 kg$$
This could also be expressed as 2 gigagrams, but I've never heard anyone use that particular unit (which might be why it sounds silly). I understand that it is impractical to use a prefix for something as large as the mass of the sun, $2\times10^{30}$ kg, but wouldn't it be more appropriate to use grams, as in $2\times10^{33}$g?
Is this simply out of convention, or is there a more logical reason?
 A: The identification of units with a standard measure happens at the anthropic level: distances, times etc units are developed from the everyday usage, thus convention is the main answer. When a discipline becomes mathematical and is concentrating at a special power of the unit, the use of  scientific "nick names" saves times and mental effort for people manipulating and discussing and communicating observations and data.
For weights,  once one gets to  atomic dimensions,  weight, in the sense of gravitational interactions, becomes irrelevant and is replaced by mass in energy related units .
So your question is about astronomic units, and it possibly is just convention that astronomers and cosmologists use kilograms, although note that often in popularizations one gets "sun masses" as a unit. In the large weights  case it is more or less convention. After all on antropic grounds we also have tons for large weights.
A: Gigagrams and megagrams are completely valid names.  But if you are often using the base unit of kilograms, it's not obvious how many of those you have.  
It is (by the rules of the BIPM) incorrect to use any additional prefix with kilograms since it already has a prefix.  $10^3kg$ may be rendered as a megagram, but (I think) with a loss of easily noting the relationship to a single kilogram.  It may not be rendered as a kilo-kilogram.
A: Divisive prefixes are fairly commomplace for the gram: I've certainly seen nanogram (ng) and microgram ($\mu$g). Multiplicative prefixes are indeed rare, and it seems that the most common is to affix to tonne (think bomb comparisons), e.g. megatonne (Mt). The tonne is accepted for use with SI units.
A: It's a weird quirk of the SI system that the base unit of mass is the kilogram, not the gram. So you'll see a lot of things expressed in kilograms.
Of course, scientists in a given field tend to standardize on certain choices of units without any regard to the SI recommendations. And this makes sense; the units you use should be the ones that make your values most understandable for the intended audience. SI is only intended as a fallback to enable unambiguous communication between groups that don't otherwise have a shared convention (especially between experimentalists and theorists). So sometimes you'll see quantities expressed in grams or tons or solar masses or whatever because that is the standard in the context you're looking at.
