MongoDB: How to add a condition on addToSet field

Question

I am using the following MongoDB aggregation to get number of articles for each day and the number of unique pinned articles :

Collection.aggregate([
    {
        $match: {
            date: { $lt: myDate },
            article: { $in: articles }
        }
    },
    {
        $group: {
            _id: { $dateToString: { format: "%Y-%m-%d", date: "$date" } },
            nbArticles: { $sum: 1 },
            distinctPinnedArticles: {
                $cond: {
                    if: { $eq: ['$isPinned', 1] },
                    then: { $addToSet: "$articleID" },
                    else: null
                }
            }
        }
    },
    {

    },
    {
        $project: {
            _id: 0,
            nbArticles: 1,
            distinctPinnedArticlesLength: { $size: { $setUnion: ["$distinctPinnedArticles"] } }
        }
    }
]

I know the condition does not work like this in MongoDB, I just want to illustrate what i want. The idea is to apply to first filter on nbArticles and then add a new filter isPinned == 1 to count the distinct articleIDs, I don't want to use facet because i am dealing with a huge number of data, so the facet can't handle a BSON with size > 16bytes.

The expected output :

[
    {
        nbArticles: 1000,
        distinctPinnedArticles: 500
    },
    {
        nbArticles: 600,
        distinctPinnedArticles: 300
    },
    {
        nbArticles: 1300,
        distinctPinnedArticles: 800
    },
    {
        nbArticles: 70,
        distinctPinnedArticles: 40
    }
]

Sample data:

[{
    _id: ObjectId("6146b91a4e98146bfba070a9"),
    article: 'article1',
    articleID: 'article_1',
    isPinned: 0,
    date: ISODate("2022-02-15T18:24:37.000Z")
}, {
    _id: ObjectId("6146b95c4e98146bfba070aa"),
    article: 'article2',
    articleID: 'article_2',
    isPinned: 0,
    date: ISODate("2021-12-01T10:45:22.000Z")
}

    , {
    _id: ObjectId("6146b9834e98146bfba070ab"),
    article: 'article1',
    articleID: 'article_3',
    isPinned: 1,
    date: ISODate("2022-01-05T08:12:59.000Z")
}

    , {
    _id: ObjectId("6146b9ae4e98146bfba070ac"),
    article: 'article2',
    articleID: 'article_4',
    isPinned: 0,
    date: ISODate("2021-11-10T14:30:45.000Z")
}

    , {
    _id: ObjectId("6146b9d54e98146bfba070ad"),
    article: 'article1',
    articleID: 'article_5',
    isPinned: 1,
    date: ISODate("2022-03-01T09:51:02.000Z")
}

    , {
    _id: ObjectId("6146ba014e98146bfba070ae"),
    article: 'article3',
    articleID: 'article_6',
    isPinned: 1,
    date: ISODate("2021-12-20T16:08:31.000Z")
}

    , {
    _id: ObjectId("6146ba2b4e98146bfba070af"),
    article: 'article2',
    articleID: 'article_7',
    isPinned: 1,
    date: ISODate("2022-01-25T12:57:09.000Z")
}

    , {
    _id: ObjectId("6146ba544e98146bfba070b0"),
    article: 'article3',
    articleID: 'article_8',
    isPinned: 1,
    date: ISODate("2022-02-10T17:39:18.000Z")
}

    , {
    _id: ObjectId("6146ba7c4e98146bfba070b1"),
    article: 'article1',
    articleID: 'article_9',
    isPinned: 0,
    date: ISODate("2021-11-30T08:20:57.000Z")
}]

Thank you in advance.

Answer 1

What about this one:

db.collection.aggregate([
   {
      $group: {
         _id: { $dateTrunc: { date: "$date", unit: "day" } },
         nbArticles: { $sum: 1 },
         distinctPinnedArticles: {
            $addToSet: {
               $let: {
                  vars: {
                     pinnedArticleID: {
                        $cond: {
                           if: { $eq: ['$isPinned', 1] },
                           then: "$articleID",
                           else: null
                        }
                     }
                  },
                  in: "$$pinnedArticleID"
               }
            }
         }
      }
   },
   {
      $set: {
         distinctPinnedArticles: {
            $filter: {
               input: "$distinctPinnedArticles",
               cond: "$$this"
            }
         }
      }
   }
])