But, why does it generally follow that hydrogen takes the place of an oxygen molecule during a hydrogenation/reduction reaction [...]
Ok, if at all, it's an oxygen atom ;)
I don't see that general trend towards complete deoxygenation, neither in catalytic hydrogenations, nor in hydride transfer reactions.
It usually takes
- rather drastical conditions
- quite some special reagents
- a particular type of derivatisation
or a combination of any of these to convert oxygen-containing compounds to alkanes.
Let's have a look at some examples:
$\ce{R1-CO-R2 -> R1-CH2-R2}$
Ketones and aldehydes may be directly reduced to alkanes in the Clemmensen reduction. The "mildest" variants use anhydrous $\ce{HCl}$ and amalgamated $\ce{Zn}$ in organic solvents.
The clasical Wolff-Kishner reduction, using hydrazine and sodium in diglyme, involves the in situ functionalization of the carbonyl and isn't a prototype for a mild reaction either. Admittedly, milder variants exist, using N-tert-butyldimethylsilylhydrazone and catalytical amounts of a Lewis acid, e.g. $\ce{Sc(OTf)3}$
Tosylhydrazones may be converted to alkanes using $\ce{NaCNBH3}$, $\ce{Na(CH3COO)3BH}$, or boranes.
Catalytic hydrogenation over Raney nickel is possible after converting the carbonyl to a dithiane; so it's actually a desulfurization.
$\ce{R1-CH2-OH -> R1-CH3}$
I'm not aware of any transformation that does not involve functionalization.
- Alkyl tosylates can be deoxygenated using $\ce{LiEt3BH}$.
- Acylation of alkohols with thiocarbonyl chlorides furnish thiocarbonates, which may be deoxygenated under using either tin hydrides ($\ce{R3SnH}$, Barton deoxygenation) or $\ce{B(CH3)3}$ in water/benzene.
- Alkanes may also be obtained from alkohols under Mitsunobu conditions ($\ce{Ph3P}$, $\ce{DEAD}$, o-nitrobenzenesulfonylhydrazine).
$\ce{R1-CH2-COOH -> R1-CH3}$
Direct transformations of carboxylic acids to alkanes typically involve decarboxylation, i.e. the loss of $\ce{CO2}$. This may be achieved either by the Barton decarboxylation, or via photoinduced electron transfer reactions with electronically excited (electron) acceptors. In the course of the latter reactions, a carboxylate is oxidized to an acyloxy radical, which then undergoes decarboxylation.
To sum it up: It's not that easy ;)