fix(NODE-5627): BulkWriteResult.insertedIds does not filter out _ids that are not actually inserted

[aditi-khare-mongoDB]

Description for NODE-5627

Before this change, on Node Driver Version 5.x, BulkWriteResult.insertedIds (1) would return an Object containing all attempted insertedIds, regardless of if any of the inserts failed. This affects BulkWrite and InsertMany operations. After this PR is merged, BulkWriteResult.insertedIds will reflect which _ids were actually inserted.

Other Notes for Reviewers

• The JIRA ticket only mentions the bug on the InsertMany operation; however, I was able to reproduce it generally for any insert operation that returns BulkWriteResult namely BulkWrite.
• BulkResult.insertedIds will still contain unsuccessful inserted _ids. The BulkResult class is not user facing.
• In this PR, unsuccessful inserts are identified through BulkResult.writeErrors list property, rather than the actual direct insert response from the server, since the batch result does not contain _id information.
• NOTE: this PR's code changes are identical to the 6.x NODE-5421 PR

Is there new documentation needed for these changes?

Yes, it needs to be communicated to users that when an insert does not succeed through a collection.bulkwrite() or collection.insertMany() operation, it will no longer appear in the error's .result.insertedIds

What is the motivation for this change?

See release highlights.

Release Highlight

Inserted ids in bulk write only contain successful insertions - 5.x

Prior to this fix the bulk write error's insertedIds field contained ids of all attempted inserts in a bulk operation.

When a bulkwrite() or an insertMany() operation rejects one or more inserts, throwing an error, the error's BulkWriteResult property will be filtered to only contain successfully inserted ids (ex: error.result.insertedIds).

Bug Description

Previous Behavior

The error's .result.insertedIds does includes all attempted inserts, regardless of if they failed or not.

New Behavior

The error's .result.insertedIds does not include unsuccessful inserts.
Namely,

• On an ordered bulk insert operation, after the first error, all future inserts fail, and therefore should not be included in insertedIds. (This is because the order would be invalid once one insert has failed).
• On an unordered bulk insert operation run on a transaction, after the first error, all future inserts fail, and therefore should not be included in insertedIds. (See here for more information.)
• On an unordered bulk insert operation that is not run on a transaction, all inserts that individually fail should be removed.

Checklist

• Ran npm run check:lint script
• Self-review completed using the steps outlined here
• PR title follows the correct format: type(NODE-xxxx)[!]: description
  • Example: feat(NODE-1234)!: rewriting everything in coffeescript
• Changes are covered by tests
• New TODOs have a related JIRA ticket

[baileympearson]

failing CI is unrelated