The sample doc schema originally defined _id as a string concatenation of metadata so that samples could be retrieved with express queries on _id. In order to be able to filter on different predicates (eg. sampleSize, samplingMethod) this metadata was also all included in the doc schema as their own fields in the doc.

Once SERVER-130444 is complete, _id will instead be a BSON object an so filtering for these predicates will be possible through the _id object alone. Remove the duplicated fields to reduce complexity. All functionality listed in the stats collection use cases must still be possible.