We are trying to use "upsert" functionality in mongoDB when storing a dataset with specific values in Id field. (If the _Id value from the dataset is found in MongoDB - update the record, if not found insert it). We use mongo-spark-connector_2.11 version 2.4.0. We set the following configuration on WriteConfig forceInsert=true, replaceDocument=true (also tried dataset.write().mode(SaveMode._Overwrite).

When trying to insert the dataset in the collection that contains id records with values from the dataset we get duplicate error (E11000 duplicate key error index: embsgdb01.xxx.$_id dup key: { : ObjectId('5d08f304ffb7a770442736c8')). My understanding is that this issue was remediated in older versions. Please let us know if anything missing from configuration or versions used.