Documentation Request Summary:

If users want to know which resume token is more recent, or want to sort based on the resume token in general, they can now do so by using the _id._data value in each document.

edit: also - for the upgrade downgrade scenarios:

• A user which opens a change stream on version 3.6 will be able to use a resume token from that stream on a 4.0 server in either feature compatibility version (FCV) 3.6 or FCV 4.0.
• A user which opens a change stream on a 4.0 binary which is in FCV 3.6 will get the old format for resume tokens, so will continue to be able to use them to resume on a 3.6 binary should they downgrade.
• Once the FCV is set to 4.0, change streams which are already open will continue using the old format, but any new change streams will use the new string format, and so will not be able to be used for resuming on a 3.6 binary.

Scope of changes:

• db.collection.watch() – _id field type change and upgrade/downgrade implications
• changeStreams - _id field type change and upgrade/downgrade implications
• Downgrade Considerations
• release notes
• db.watch() just mention type in the _id field description
• Mongo.watch() just mention type in the _id field description

Impact to other docs outside of this product:

MVP:

Resources:

Engineering Ticket Description:

This will enable clients to sort on the value, without the interesting and surprising sorting semantics of the BinData type.