How to import large xml file into PostgreSQL (sqlalchemy)?

Question

Python

Now I do like this:

class XmlHandler(ContentHandler):

    def startElement(self, name, attrs):
        if name == 'Objects':
            print('start')
        elif name == 'Object':
             self.obj = attrs['param]'

    def endElement(self, name):
        if name == 'Objects':
            print('end')
        elif name == 'Object':
             self.obj.save()

xml.sax.parse(xml_path, XmlHandler())

Which, of course, is very slow.
And related questions:
1. How to do it faster?
2. How to calculate which operations of this particular section of code take the most CPU time?

Answer 1

At least I'm wrong that I did not show the entire implementation of the method.
At the endElement at the end I had gc.collect() The
profiler suggested that the first problem was
pix.toile-libre.org/upload/original/1503120966.png
After
pix.toile-libre.org/upload/original/1503121255. png
search for an object in the database

class Obj(Model):
    id = Column(UUID, primary_key=True)

class XmlHandler(ContentHandler):
    def __init__(self):
        pass

    def startElement(self, name, attrs):
        if name == 'Objects':
            print('start parse')
        elif name == 'Object':
            aoid = attrs['ID']
            self.obj = Obj.query.filter_by(id=id).first() # это 2-ое узкое место
            if not self.obj:
                self.obj = Obj()


    def endElement(self, name):
        if name == 'Objects':
            print('end')
        elif name == 'Object':
        	self.obj.save()
        gc.collect() # это было 1-ое узкое место

xml.sax.parse(xml_path, XmlHandler())

What can be done about it?