Web site structure search results table shows wrong # of sites

[noam.bernstein]

I was looking for TiO2 structures, and I found the following into for tetragonal structures in the overview table:

(note the number of sites for the two tetragonal structures, mp-390 and m-2657)

Now, when I check mp-390 (# sites 6) and mp-2657 (# sites 12) I get

12 sites for mp-390, and

6 sites for mp-2657

I assume the overview table is being generated from the underlying data, so I can’t imagine how this is happening, but it looks like a bug.

[tschaume]

@noam.bernstein thanks for reaching out. I think this forum post should answer your question.

[noam.bernstein]

I don’t think it does. First, that forum post is mostly about consistency between the web interface and the API, and I think that consistency between two web interface views is a lot more important.

Second, I can see how one view providing the conventional cell and the other the primitive cell would lead to different site numbers, but I was under the impression that the conventional cell is always >= the primitive cell. If one of those views (table vs. individual structure) was primitive and the other conventional, the conventional # sites should always be >= the primitive number. But in my example in one case the structure view is >, and in the other it’s <. I can’t think of any consistent rule as to what each view is showing that would lead to those values. If there is such a rule, could you explain it?

[tschaume]

Yes, you’re right. Good catch! We’re investigating and will get back to you.

[tschaume]

@noam.bernstein it turns out that this is indeed an issue with our current database version. It will be fixed with our (imminent) upcoming database release. Thank you for your patience.