1.3.1pre11: Duplicate star

[ajtribick]

After analysing the stars.dat file supplied with pre11 (as part of preparation for a star database merger), I found that the star HIP 118322 is entered twice.

To check that this was not a glitch of my code, I entered Celestia and did a Go To HIP 118322, and sure enough the lens flare was much brighter than it should be, indicating the presence of two stars in the same location.

[JackHiggins]

Lens flare?!!

[selden]

Jack,

Now, now, don't get all excited. I'm sure Chaos is just referring to the flare that Celestia has always shown around stars when you're close to them. e.g.

[JackHiggins]

Now, now, don't get all excited.

Phew... Thought we had a great new feature on our hands there for a second!

[granthutchison]

If you check the Star Browser when you're next to this star (Eps Tuc), you also find that it appears twice there. I reported this a while back but dumbly assumed it was some error in the way the Celestia code handled stars.dat (since this star is the last one in the database), rather than a simple duplication in the dataset itself.
Just checking with Chris that the final-entry duplication isn't serving some arcane coding purpose, and then I'll commit a correction for this.

Grant

[granthutchison]

Thought we had a great new feature on our hands there for a second! :)

Oh my goodness, no! I hope we never have lens flare in Celestia. Annoys the hell out of me.

Grant

[ajtribick]

On investigation into the really huge 50MB star database, I find two duplicated star systems:

HIP 118322 (as in the star database supplied)
A star with the number 33672726 in its Hipparcos catalog field, I presume this is in the Tycho catalog, but TYC 3367-272-6 does not appear to exist.

Irritatingly I accidentally closed the window and it takes about ten minutes for my program to read in the data...

EDIT: The star is TYC 2726-3367-0 (thanks to Selden for pointing to the page about fictitious Hipparcos numbers for Tycho stars)

[marc]

I confirm that duplicate star, I use the hip number for the primary key in the mh database. here is the log file.

Code: Select all

Starting generation process
error dropping star table  Unknown table 'star'
Duplicate entry '118322' for key 1

chaos syndrome,
Have you had a try of my mysql addon for celestia?
If your willing to learn a bit of SQL it is very useful for inspecting stars.dat from within celestia.

here are the links.
http://celestiaproject.net/forum/viewtopic.php?t=2518
http://mostlyharmless.sourceforge.net/s ... index.html

The next version (0.1.4) of my mysql addon has the ability to output stars.dat and starnames.dat files. It makes it easy to add names to stars, make any modifications, and even merge catalogs. (all from within celestia )
Ive got a few more changes and a little tutorial on basic editing to write before i release it, but it should be out soon.

[ajtribick]

chaos syndrome,
Have you had a try of my mysql addon for celestia?
If your willing to learn a bit of SQL it is very useful for inspecting stars.dat from within celestia.

Hmmm... looks interesting, but I can't face a 3MB download right now... dial-up connections are such fun

Anyway, in examining the really large extended stars database, it turns out that Epsilon Tucanae is represented twice, but has different spectral class fields, one is reported as B5V, the other as B9IV.

As far as I have been able to tell, Eps Tuc is a B9 star, so I think the B5V entry is going to get chopped out.

[granthutchison]

I've commited the revised version of stars.dat, though it'll probably take a while to show up in the CVS tree.

Grant

[HB]

About a year ago, the existance of this double star has been mentioned in this forum.
At that time I've build stars.dat myself with "Buildstardb.exe". Analysing the stars.dat file after building the datbase has sorted out that this double star is caused by this buildstardb.exe program.
I've been spitting and bugging around in the source code of this program and found the cause of this problem. I've sent this debugged sourcecode to Chris for uploading to CVS. (together with another program for analyzing this database)Unfortunatly this improved version of buildstardb is still not available for the users of Celestia.
So this double star has nothing to do with the version of Celestia.

[granthutchison]

So this double star has nothing to do with the version of Celestia.

My understanding is that buildstarsdb was used as a one-off to create the original stars.dat, which has subsequently been edited in the various Celestia versions to correct various errors like the one above. Certainly it's possible to edit stars.dat and have the corrections carry over into subsequent releases of Celestia.

Grant