It will be slow. OData is not a bulk API. It is designed so that clients can access individual objects and move between them, in most cases break pages into some filtered lists.
Retrieving the entire dump through OData will not make anyone happy. The owner of the OData API will need to investigate who does all these nightly crawls on his API and discovers that you are the one and probably disconnected you. On the other hand, you will find that OData is not an efficient mass transport format, but marshaling HTTP encoded objects back and forth is not the best way to spend your bandwidth. And crawling the entire database every time, unlike just detecting the deltagrams from the last crawl, will only work until the database reaches such a critical size S that the update takes longer than the frequency you combine!
Also, if this is not your data, it is highly likely that terms of use for the OData API explicitly prevent such massive crawls.
Get a data dump, archive it and copy using FTP.
source share