The JSON snapshot generated for Metadata Plus users is around 20 GB smaller than last months. Uncompressed, it is larger than last months and contains all of the latest metadata.
So, what happened? We recently made an upgrade which affected pagination. As part of the new implementation of the cursor parameter, we forced an ordering of the results. The process we use to generate the snapshot makes use of pagination using cursors, and it turns out that the ordering meant that each individual file could be compressed at a higher rate. We’ve seen this kind of effect before — it happened when we put this year’s public data file and ordered by DOI.
The contents of the JSON snapshot is exactly the same as it would have been using the old process, but the contents of each individual file within the snapshot will be different. Of course, if you see something that looks off please leave a comment or get in touch with our support team.