The Csus2 chord implies that the 2nd (the D note) is suspended from a previous harmony. Thus the example is just to show how it can happen and that it can sound ok. Ultimately all of this stuff has its origins in part writing and over years rules have been written and broken and re-written. Bach said "no consecutive fifths or octaves between any parts". I doubt that the average Big Band would get through 8 measures with that rule...

It is correct that a suspension is dissonant (ie: discordant, or unstable) and requires resolving. Of course the most discordant in the examples is the Bb in the dominant 7th (C7) resolving to the the third of the tonic (A note in the F chord). It's a similar effect to a suspension (just a stronger pull). The suspension could be held over a longer period, to create more tension, maybe even never resolving before the progression moves on.

I think your problem is assuming that the examples are for auto play chord. Of course (without writing it out in full) the root bass note is always assumed to be sounding. You are right of course that the whole basis of the harmony alters depending on the root. An example being that of the Tritone (the augmented 4th or diminished 5th) for example Bb and E. This makes the most dissonant or discordant of harmonic intervals which might be assumed to resolve to A and F (C7 to F), (assuming a bass note moving from C to F) In this case the Bb is the 7th of the dominant 7th harmony and the E the leading note. Change the first bass note to F# and this is very clearly F#7 which in a perfect cadence resolves to B (In this case it would be A# and E, the E now being the 7th of the dominant and the A# being the leading note (The 7th would fall to D#, the mediant of the tonic, and the A# would rise to the root of B).

But as to whether this chord (Csus2) is truly Em7... it cannot be because the root of the chord is... C. If it said Csus2/E then it could possibly be Em7 by another name. Especially that if you need to omit a note from a harmony the fifth is the safest to omit without altering the effect. But it says clearly Csus2, so this is a C chord. All this proves is that although chord symbols are very useful for speed reading and of course APC, they are a musical short hand.