Inverted Joins

Suppose that the purpose of the join is to retain all images in the primary collection that are not in the secondary collection. You can perform this type of inverted join using ee.Join.inverted().

Code Editor (JavaScript)

// Load a Landsat 8 image collection at a point of interest.
var collection = ee.ImageCollection('LANDSAT/LC08/C02/T1_TOA')
    .filterBounds(ee.Geometry.Point(-122.09, 37.42));

// Define start and end dates with which to filter the collections.
var april = '2014-04-01';
var may = '2014-05-01';
var june = '2014-06-01';
var july = '2014-07-01';

// The primary collection is Landsat images from April to June.
var primary = collection.filterDate(april, june);

// The secondary collection is Landsat images from May to July.
var secondary = collection.filterDate(may, july);

// Use an equals filter to define how the collections match.
var filter = ee.Filter.equals({
  leftField: 'system:index',
  rightField: 'system:index'

// Define the join.
var invertedJoin = ee.Join.inverted();

// Apply the join.
var invertedJoined = invertedJoin.apply(primary, secondary, filter);

// Print the result.
print('Inverted join:', invertedJoined);

The output should look something like:

Image LANDSAT/LC08/C02/T1_TOA/LC08_044034_20140403 (17 bands)
Image LANDSAT/LC08/C02/T1_TOA/LC08_044034_20140419 (17 bands)

The inverted join contains the images from April 3 and April 19, indicating the images that are present in the primary collection but not in the secondary collection.