Why do we get a diffraction pattern with a slit whose width is not comparable to the wavelength of the incident light? In single slit experiments with sodium light whose wavelength is of order micrometer and slit width of order of hundreds of micron, we still observe the diffraction pattern. How do we explain this?  
 A: The often heard claim that  

"diffraction only happens when the slit size is comparable to the wavelength"

(or variations thereon) is mis-stated in a overly broad way for ease of rattling it off. 
A more complete understanding comes from examining how the results of the diffraction experiments vary as a function of the feature size. The angles involved (angular location of the $n$th maxima or the $m$th minima) go by the ratio 
$$\frac{\text{wavelength}}{\text{linear feature scale}} \,,$$
for many diffractive experiments and demos.
When this ratio gets small the fringes are very close together (unless you project over very long distances). 
As a result, the visible fringes are easily washed out when the feature scale is too big. For instance, use a broad band source and too large a slit size or spacing and the overlapping of different order maxima of various wavelengths can end up looking like a slight ripple of color that is not easily pulled apart into distinct bits, unless you project them a long way, indeed. The effects are still there, but limitation of the human visual system are getting in the way of your seeing them.
If the angular scale of the features ever drops below the angular dispersion of the beam from your source the pattern will be washed out at all projection distances.
Then there is the matter of intensity. Projecting a long way to increase the linear size of the projected pattern means that you pay the usual cost in intensity. If you start with a relatively dim source, you won't be able to see the pattern with the naked eye.
But a carefully constructed experiment can still demonstrate diffractive physics with large features. See, for instance, my answer to a question about making diffraction grating with a laser printer. The scale of my slits is much larger than the wavelength, but by using a narrow-band source (a laser in this case, but the sodium D lines from your lamp are almost as good) and projecting over eight meters I get a distinguishable pattern. The pattern is quite lost if projected over less than one meter because the size of dots is larger than their spacing.
So, it might be a more accurate to say something like

"The results of diffraction are easily observed when the size of the slit is comparable to the wavelength."

But those words also carry the possibility of misinterpretation when given to a beginner.
A: You always get a diffraction pattern, no matter the slit or object size. The issue becomes: "Is it focused well enough to be observable?" 
You get a nicely viewable diffraction pattern under finely tuned circumstances where the slit is small with respect to the wavelength, but even a baseball creates a diffraction pattern interfering with itself, however at that scale any observable features are washed out and you would have to go to the math to do your "observing."
