MongoDB Compass Export Pipeline to Language with other style

Question

In MongoDB Compass, when I choose to export the Aggregation Pipeline to Java, I get something like this:

Arrays.asList(new Document("$group", 
    new Document("_id", "$loginTime.seconds")
            .append("loginTime", 
    new Document("$min", "$loginTime.seconds"))))

While this seems to be correct, I'd like to know how I can generate the equivalent expression:

List.of(Aggregates.group("$loginTime.seconds", Accumulators.min("loginTime", "$loginTime.seconds)));

Obviously the latter is more straight-forward and declarative enough to understand on first sight. But MongoDB Compass does not offer any option for that. Why.

Answer 1

I researched a bit based on the comment from @prasad_ It seems that there are related issues on the mongoDB issue tracker see https://jira.mongodb.org/browse/COMPASS-4695

I also think that this would be a very nice feature since longer aggregations are just unreadable and if we need to adapt these we have to re-engineer them quite often.

Have you thought about saving your aggregation stage as a view in the database? Then the pipelines are stored in the mongodb and you can query it like any other collection.

Other than that I thought about exporting the pipeline as js / json / or raw pipeline (I prefer js/json for auto-formatting) and then parse it from a resource folder.

Of course this would mean that you have to somehow come up with a mechanism to replace dynamic parts of the query which could lead to security issues probably, but maybe you can break down parts that are not dynamic.

    Aggregation aggregation =
    newAggregation(
        CustomDocumentAggregationOperation.of(
            matchByIdAndOtherValues(objectId, otherValues),
            lookupSomeDataOfOtherCollection(),
            unwindSomething(),
            projectToSomething()));

I adapted this pattern within my repositories so that I can see at a first glance what the aggregation is doing. Hidden behind these static functions is a bunch of bson.Documents and I try to avoid going there...

This is how I assemble multiple Documents to one AggregationOperation

 public class CustomDocumentAggregationOperation implements AggregationOperation {
  private final Document document;

  public CustomDocumentAggregationOperation(Document document) {
    this.document = document;
  }

  @Override
  public Document toDocument(AggregationOperationContext aggregationOperationContext) {
    return aggregationOperationContext.getMappedObject(document);
  }

  public static AggregationOperation of(Document document) {
    return new CustomDocumentAggregationOperation(document);
  }

  public static List<AggregationOperation> of(@NonNull Document... documents) {
    return Arrays.stream(documents)
        .map(CustomDocumentAggregationOperation::of)
        .collect(Collectors.toList());
  }