Sample

Gets a random sample from the collection, by a specified number of docs (n).
Useful to...

• do initial analysis on a dataset
• do sampling on result set
• fetch docs for random-user-searching
• seed some random obj for computation

Conditions

When... n < 5% of the collection AND

100 docs in the collection AND sample is the first stage, a pseud-random cursor selects docs for returning.
When... those conditions are NOT 100% met, an in-memory random sort happens. This has the 100mb restriction.

example

db.nycFacilities.aggregate([
  {
    $sample: { size: 200 }
  }
])

• collection has > 100 docs
• sample size is > 5% of docs
• sample stage is FIRST of the pipeline
• ... pseudo-random operation will apply