I'm trying to emulate the following MongoDB shellcode:
db.collection.find( { $and: [ { $or: [ { document: { field: "X" } }, { field: "X" } ] }, { _id: ObjectId("X") } ] } );
This is what I've tried (with new MongoDB-C-Driver):
bson_init(&query);
bson_append_document_begin(&query, "$and", 4, &and);
bson_append_oid(&and, "_id", 3, oid);
bson_append_document_begin(&and, "$or", 3, &or);
bson_append_utf8(&query, "field", 5, "X", 1);
bson_append_document_end(&and, &or);
bson_append_document_begin(&and, "$or", 3, &or);
bson_append_utf8(&query, "document.field", 14, "X", 1);
bson_append_document_end(&and, &or);
bson_append_document_end(&query, &and);
collection = mongoc_client_get_collection (client, "db", "collection");
cursor = mongoc_collection_find(collection, MONGOC_QUERY_NONE, 0, 1, 0, &query, NULL, NULL);
if(mongoc_cursor_next(cursor, &doc)){
printf("> Field found\r\n");
}
Thank you in advance.
Best regards.
The libbson imperative API for creating nested documents is a little tricky and unfortunately you've stumbled into one of the easy pitfalls. Once a subdocument has been opened with bson_append_document_begin or bson_append_array_begin, you must not write to it until a corresponding _end() call has been performed. In this case, you have append_utf8() calls in your "or" documents that write to "query".
For an easier approach to bson composition, consider using the BCON api, which provides a more declarative syntax with minimal overhead:
using the bcon api might also have given you a hint that you weren't quite replicating what you thought you were.
To produce the bson you were looking at in the shell:
You can also use the bson_as_json() function to stringify a bson document into json, which should make it easier to see what object you've constructed
The relevant documentation: http://api.mongodb.org/libbson/current/