Datasets:

ArXiv:
Tags:

How to join documents and quality signals when downloading directly

#19
by tgshdyfuhuf - opened

I've downloaded the documents and quality signals separately with wget
Surprisingly, I don't see which primary key I can use to re-join the two data sources. How can I do this?

Together org

Hi @tgshdyfuhuf

Each quality signal has a field "id" which points to the idx of the corresponding record in the documents files. For example the id 2023-06/0000/de_head.json.gz/262 means that the signal is associated with document 262 (0-indexed) in the documents file 2023-06/0000/de_head.json.gz.

@mauriceweber safe assumption to make that each signals file will only contain ids pointing to the same namespaced documents file
ie. signals 2023-06/0000/de_head.json.gz will only contain signals for document 2023-06/0000/de_head.json.gz

Together org

yes exactly -- there is a one-to-one correspondence between rows in the signals file and rows in the documents file

Sign up or log in to comment