3

In the mongocxx API, Collection.aggregate() expects a pipeline object in order to run an aggregate pipeline query. This means constructing the query by using the Pipeline class. Such as:

    mongocxx::pipeline p{};
    p.match(make_document(kvp("items.fruit", "banana")));
    p.sort(make_document(kvp("date", 1)));
    auto cursor = db["sales"].aggregate(p, mongocxx::options::aggregate{});

Is there a way to run an aggregate pipeline query in mongocxx by passing in a string? I'm not looking to construct the query using a mongocxx object, but running the query as a string.

For example:

    db["sales"].aggregate("[{"$match": {  ... }}"]

where "[{"$match": { ... }}" is a pipeline aggregate query of type std::string.

anon9928175234
  • 105
  • 1
  • 8
  • You can parse a string into a bson document instead of building one up, is that what you're asking? – rustyx Aug 17 '18 at 18:24
  • Yes, I'm looking to have the entire query as a string which I can run using mongocxx. – anon9928175234 Aug 17 '18 at 20:01
  • Not the entire string, but the pipeline components (match, sort, etc.) you can parse as JSON fragments using `bsoncxx::from_json`. – rustyx Aug 17 '18 at 20:17
  • No, there is not currently a way to construct such a thing. However it may be very easy to add. Would you mind filing a ticket? https://jira.mongodb.org/projects/CXX – acm Aug 17 '18 at 20:58

1 Answers1

1

Yes you can use run_command of mongocxx::database

bsoncxx::builder::basic::document command_document{};

command_document.append(kvp(
"eval",
"function(username) {"
"return db.users.findOne( { username : username } );"
"}"));

command_document.append(kvp("args", [&](sub_array child) {
child.append(username);
}));

auto doc = db.run_command({command_document});

This is a simple example of using string function with arguments to run on Mongodb by mongocxx now you can use it for any command that you want. Is this that you need ?

FONQRI
  • 368
  • 4
  • 13