Problems importing large file in Core Data with background thread - swift

Question

I am running into some trouble importing a large .csv file in OS X using Core Data on a background thread.
My simplified data model is a Dataset which has a to-many relationship to many Entries. Each entry is a line in the .csv file (which has a bunch of attributes that I'll leave out for brevity). Following what I have read about efficiently importing lots of data, along with how to make a progress indicator work properly, I have created a managed object context for the purposes of going the import.
I am having trouble with two things:
I need to hang on to a reference to the new Dataset at the end of the import, because I need to select it in the popup. this will be done in the main thread, but for efficiency (and to make my NSProgressIndicator work) the new dataset is created on the background thread, with the background MOC.
From what I read, batching the import, so that the background MOC saves and resets, is the best way to stop the import from eating up too much memory. That does not turn out to be the case so far - it looks like gigs of memory is being used even for files in the tens of megabytes. Also, once I reset my import MOC, it cannot find the data for the inDataset, and so cannot create the relationship between all subsequent Entries and the Dataset.
I've posted the simplified code below. I have tried refreshObject:mergeChanges, without good results. Can anyone point me at what I am doing wrong?
let inFile = op.URL
dispatch_async(dispatch_get_global_queue(priority, 0)) {
//create moc
let inMOC = NSManagedObjectContext()
inMOC.undoManager = nil
inMOC.persistentStoreCoordinator = self.moc.persistentStoreCoordinator
var inDataset : inDataset = Dataset(entity: NSEntityDescription.entityForName("Dataset", inManagedObjectContext: inMOC)!, insertIntoManagedObjectContext: inMOC)
//set up NSProgressIndicator here, removed for clarity
let datasetID = inDataset.objectID
mocDataset = self.moc.objectWithID(datasetID) as! Dataset
let fileContents : String = (try! NSString(contentsOfFile: inFile!.path!, encoding: NSUTF8StringEncoding)) as String
let fileLines : [String] = fileContents.componentsSeparatedByString("\n")
var batchCount : Int = 0
for thisLine : String in fileLines {
let newEntry : Entry = Entry(entity: NSEntityDescription.entityForName("Entry", inManagedObjectContext: inMOC)!, insertIntoManagedObjectContext: inMOC)
//Read in attributes of this entry from this line, removed here for brevity
newEntry.setValue("Entry", forKey: "type")
newEntry.setValue(inDataset, forKey: "dataset")
inDataset.addEntryObject(newEntry)
dispatch_async(dispatch_get_main_queue()) {
self.progInd.incrementBy(1)
}
batchCount++
if(batchCount > 1000){
do {
try inMOC.save()
} catch let error as NSError {
print(error)
} catch {
fatalError()
}
batchCount = 0
inMOC.reset()
inDataset = inMOC.objectWithID(datasetID) as! Dataset
// fails here, does not seem to be able to find the data associated with inDataset
}
}// end of loop for reading lines
//save whatever remains after last batch save
do {
try inMOC.save()
} catch let error as NSError {
print(error)
} catch {
fatalError()
}
inMOC.reset()
dispatch_async(dispatch_get_main_queue()) {
//This is done on main queue after background queue has read all line
// I thought the statement just below would refresh mocDataset, but no luck
self.moc.refreshObject(mocDataset, mergeChanges: true)
//new dataset selected from popup
let datafetch = NSFetchRequest(entityName: "Dataset")
let datasets : [Dataset] = try! self.moc.executeFetchRequest(datafetch) as! [Dataset]
self.datasetController.addObjects(datasets)
let mocDataset = self.moc.objectWithID(datasetID) as! Dataset
//fails here too, mocDataset object has data as a fault
let nDarray : [Dataset] = [mocDataset]
self.datasetController.setSelectedObjects(nDarray)
}
}