Sylvester to Alexey Prokudin

In a couple of cases we have a bin with no meaningful measurement. These bins are almost exclusively restricted to outside of the recommended z-range of 0.2-0.8. Generally I marked these bins by zeroing out the entire line in the data files.

Internally, I encode the reason why a bin is marked bad in an error code, which I store as a negative value for the multiplicity, allowing me to carefully check the data integrity.

When I first put the multiplicity files online, I accidentally uploaded a version that did not have all the error codes purged out of the data files. This issue already came up a couple of months ago, and has since been fixed in the database.

My advice would be to either download the new files, or make your code ignore all bins with a negative multiplicity value.