After knowing that self ballasted mercury vapor and metal halide lamps were made, I have noticed there is apparently no such thing as a self ballasted high pressure sodium lamp or a self ballasted low pressure sodium lamp. I wonder why such lamps were never marketed.

Inefficiency I would say?

By the "selfballasted MH" you mean the modern 20W CMH with integrated electronic ballasts?

Otherwise with a simple resistive ballast works only the MV, no other. The reason is the others have way less stable arc.
Plus there is the motivation: Selfballasted MV was designed as a better light quality alternative to the MV, yet a bit more efficient and mainly with higher CCT than incandescents.
At the MH era higher CCT and light quality were not any motivation anymore, as MH's allowed all at an acceptable level without the filament. Plus as mentioned above, the arc instability made the ballast quite complex, negating the simplicity of just a series tungsten filament in MV's. The only MH with incandescent based ballast was the GE halarc experiment and it was a sales flop.
Only the modern electronic allowed to design a small and cheap enough high efficiency ballast to be integrated with the arctube into an integrated lamp product, which then allowed a simple use instead of an incandescent yet offered the high efficacy of CMH. But due to heat limitation, this was limited to below 35W rating.


HPS use is motivated by high efficiency (including high efficiency of optical coupling to the optics) where light quality does not matter. So again beside the technical difficulties to make the arc working with a resistive ballast, the losses of the incandescent quite negate the main purpose why to use HPS in the first place.

The electrical characteristics of the HPS arc are entirely different than HPM and MH.  The plasma temperature is lower, and when operated on a resistive ballast the reignition peaks of the discharge are too high to sustain with a normal ballast filament.  It would be necessary to make the HPS arc tube with a much lower RMS voltage than usual, which in turn means more voltage and power dissipated across the filament, and that in turn means a bigger efficacy decrease than for HPM self-ballasted lamps.  About 15 years ago we did develop such a lamp at Sylvania but it was not marketed.  The voltage waveform of an HPS arc is rather more favourable for the neon-argon filled arc tubes, and with those it was possible to keep the arc burning with a relatively efficient tungsten filament.  Since those lamps were in reflector bulbs and intended for horticultural lighting applications, the added deep red radiation from the filament was not entirely useless.  But even so, they were still rather too sensitive to mains voltage fluctuations.

Some years before that, Ted Glenny from Philips UK filed for a patent on a self-ballasted HPS lamp in which he arranged for an unusually high mole ratio of mercury in the arc tube, which made the HPS arc less efficient than normal but claims to have facilitated the self-ballasted design.  See https://depatisnet.dpma.de/DepatisNet/depatisnet?action=pdf&docid=EP000000017280A1&xxxfull=1

There was another Philips patent that involved placing a small electronic voltage doubler device inside the lamp cap, to overcome the problem of the reignition voltage.  I cannot find that one right now.

For LPS lamps, these problems would be exacerbated by their relatively high ignition voltage.  Even in normal use they are run on transformer ballasts or with a separate ignitor.  Due to the additional volt drop across the filament, an even higher open circuit voltage would be required, and that would make the transformers less efficient.  And if you still need a transformer anyway, there is little point to make a self-ballasted lamp.  During a visit to Philips' Hamilton plant, I did however see some electronically-ballasted SOX-E 18W lamps.  Those were basically using the ballast and plastic base of a compact fluorescent lamp.

I wonder why self ballasted high pressure sodium lamps even with integrated electronic ballasts also never made it into the market too.

Again: The main and only point of HPS was to provide the cheapest light for long time use, even when it means lower light quality.
Making the rather expensive part, that can be made to last (tge classic choke), disposable isn't going to make the light cost low.

The integrated MH's were intended to go after the high light quality retail display market, where along the light quality the variability of the installation is needed. Znd because that market is used to rather higher prices, it was not a problem to budget in the disposable electronic ballast, when it allows the higher efficiency and longer life of MH vs halogens while maintaining the compact form and light weight (needed for the variability)...