snap-stanford / ogb

Benchmark datasets, data loaders, and evaluators for graph machine learning
https://ogb.stanford.edu
MIT License
1.89k stars 397 forks source link

Mismatching between xyz in order and edge index #309

Closed jiaqingxie closed 2 years ago

jiaqingxie commented 2 years ago

Hi. Weihua. We found this issue: for example in an xyz file, the data might be in this format: C 0 0 0 C 0 1 1 C 0 2 2 N 1 0 1 Si 1 0 2 with this index order: 0, 1, 2, 3, 4 It is just an assumption. And we just plotted them in a 3d space with each index. But we found that it might not be matching with the order that we have printed for example C, N, C, C, Si but not C, C, C, N, Si which is corresponding to the edge index.

We also figure out that the order of C is confused, which means that in the edge index matrix, the index of the first C is 0 but is actually the index 1 in the original xyz file.

We don't know how to deal with it.

weihua916 commented 2 years ago

Hi! I suggest you use the newly-released SDF format: instructions here.