Update 4. Aggregation Pipelines: Let the Server Do It For You.md

Author: ortizfram

Introduction to MongoDB in Python/4. Aggregation Pipelines: Let the Server Do It For You.md

@@ -97,3 +97,31 @@ pipeline = [
 print(list(db.laureates.aggregate(pipeline)))
 ```
       [{'_id': None, 'n_prizes_total': 27}]
+## 🦍 Gap years, aggregated
+- [x] Make the $group stage output a document for each prize year (set "_id" to the field path for year) with the set of categories awarded that year.
+- [x] Given your intermediate collection of year-keyed documents, $project a field named "missing" with the (original) categories not awarded that year. Again, mind your field paths!
+- [x] Use a $match stage to only pass through documents with at least one missing prize category.
+- [x] Finally, add sort documents in descending order.
+```py
+from collections import OrderedDict
+
+original_categories = sorted(set(db.prizes.distinct("category", {"year": "1901"})))
+pipeline = [
+    {"$match": {"category": {"$in": original_categories}}},
+    {"$project": {"category": 1, "year": 1}},
+    
+    # Collect the set of category values for each prize year.
+    {"$group": {"_id": "$year", "categories": {"$addToSet": "$category"}}},
+    
+    # Project categories *not* awarded (i.e., that are missing this year).
+    {"$project": {"missing": {"$setDifference": [original_categories, "$categories"]}}},
+    
+    # Only include years with at least one missing category
+    {"$match": {"missing.0": {"$exists": True}}},
+    
+    # Sort in reverse chronological order. Note that "_id" is a distinct year at this stage.
+    {"$sort": OrderedDict([("_id", -1)])},
+]
+for doc in db.prizes.aggregate(pipeline):
+    print("{year}: {missing}".format(year=doc["_id"],missing=", ".join(sorted(doc["missing"]))))
+```