Workaround dictionary sorting

Question

Through quick research I found out that there is no way to truly sort a dictionary. However there are workarounds and I used the one suggested here by user Malik.

I ran this test and got:

>>> test={"im2":1,"im4":2,"im3":-1,"im1":5}
>>> {x:test[x] for x in sorted(test)}
{'im1': 5, 'im2': 1, 'im3': -1, 'im4': 2}

which is exactly the behaviour I'm looking for. Any help or suggestion is greatly appreciated.

However, when I try to incorporate it into my code, nothing changes. Here's the code:

def loop_solution(filenames):
    mydic={'im'+str(index): import_file_astro(filename) for index, filename in enumerate(filenames, start=1) if filename is not None}
    print mydic
    mydic_upd={'im'+str(i):mydic.values()[0] for i in range(1,5) if 'im'+str(i) not in mydic}
    mydic=dict(mydic, **mydic_upd)
    print mydic
    mydic={y:mydic[y] for y in sorted(mydic)}
    print mydic
    return
loop_solution([None,'CO21_m100_100_final',None,'CO43_m100_100_final'])

and the output:

{'im2': <astropy.io.fits.hdu.image.PrimaryHDU object at 0x11b6dc690>, 'im4': <astropy.io.fits.hdu.image.PrimaryHDU object at 0x11b7461d0>}
{'im1': <astropy.io.fits.hdu.image.PrimaryHDU object at 0x11b6dc690>, 'im3': <astropy.io.fits.hdu.image.PrimaryHDU object at 0x11b6dc690>, 'im2': <astropy.io.fits.hdu.image.PrimaryHDU object at 0x11b6dc690>, 'im4': <astropy.io.fits.hdu.image.PrimaryHDU object at 0x11b7461d0>}
{'im1': <astropy.io.fits.hdu.image.PrimaryHDU object at 0x11b6dc690>, 'im3': <astropy.io.fits.hdu.image.PrimaryHDU object at 0x11b6dc690>, 'im2': <astropy.io.fits.hdu.image.PrimaryHDU object at 0x11b6dc690>, 'im4': <astropy.io.fits.hdu.image.PrimaryHDU object at 0x11b7461d0>}

I'm at a loss of what could be the problem.

Have you tried adding `return mydic` at the end of your function and then `mydic = loop_solution([...........])` outside of your definition? — vash_the_stampede, Sep 11 '18 at 13:04
@Aran-Fey Yes but the workaround appears to follow the same logic. — George, Sep 11 '18 at 13:07
@George doesn't work in python<3.6 though. As dexcribed in the linked duplicate, the insertion order is not preserved. Only in 3.7 does this become a language feature — FlyingTeller, Sep 11 '18 at 13:11
@FlyingTeller Shouldn't then a problem arise in the first snipet of code I posted? — George, Sep 11 '18 at 13:13
@vash_the_stampede This seems to work but I need it to happen inside the function. — George, Sep 11 '18 at 13:16
@George I don't see why. The docs say that the key order is "arbitrary", which is not the same as "always scrambled up". Might be that insertion order is preserved in some cases (probably dependent on keys, values and length), but you cannot depent on it, neither should you. — FlyingTeller, Sep 11 '18 at 13:19
I also don't see why you need that. Anytime you loop over your dict, you could just do `for key in sorted(my_dict.keys())` and be save — FlyingTeller, Sep 11 '18 at 13:19
Okay then `def loop_solution(filenames, mydic)` pass a empty `mydic` into the function? — vash_the_stampede, Sep 11 '18 at 13:21

Workaround dictionary sorting

0 Answers0