Here's a graph of the impedance vs frequency behavior of a "randomly" selected SMT ferrite bead inductor:
You can see that the impedance of the component varies widely over frequency.
Also, the real part of the impedance (R) dominates the behavior over most of the frequency range, and the imaginary part (reactance, X) dominates only at relatively low frequencies.
When the datasheet tells you the impedance is 100 ohms at 100 MHz, that is just giving you just one point on this curve. It's technically only telling you the magnitude of the total impedance, but in most cases this will be almost entirely real impedance (resistance).
The device I showed the plot of has a listed typical impedance of 120 ohms at 100 MHz, which you can see agrees with just that one place on the curve.
Some devices might still be increasing in impedance at 100 MHz (like this one), while others could be already past their peak and in the declining part of the curve (as happens above about 1000 MHz in my example part).