From a206e777af4378c2bb191d15064a62016bede987 Mon Sep 17 00:00:00 2001
From: "Albert (AJ) St. Aubin" <astaubin@edx.org>
Date: Thu, 27 Aug 2020 14:19:58 -0400
Subject: [PATCH] Write a migration to clean up and secure duplicate
UserDemographics entries
[MICROBA-551]
---
.../0002_clean_duplicate_entries.py | 48 +++++++++++++++++++
.../migrations/0003_auto_20200827_1949.py | 20 ++++++++
.../core/djangoapps/demographics/models.py | 2 +-
3 files changed, 69 insertions(+), 1 deletion(-)
create mode 100644 openedx/core/djangoapps/demographics/migrations/0002_clean_duplicate_entries.py
create mode 100644 openedx/core/djangoapps/demographics/migrations/0003_auto_20200827_1949.py
diff --git a/openedx/core/djangoapps/demographics/migrations/0002_clean_duplicate_entries.py b/openedx/core/djangoapps/demographics/migrations/0002_clean_duplicate_entries.py
new file mode 100644
index 00000000000..bfadadf061b
--- /dev/null
+++ b/openedx/core/djangoapps/demographics/migrations/0002_clean_duplicate_entries.py
@@ -0,0 +1,48 @@
+import logging
+
+from django.conf import settings
+from django.db import migrations, models
+
+log = logging.getLogger(__name__)
+
+
+def _clean_duplicate_entries(apps, schema_editor):
+ """
+ This method finds all the duplicate user entries in the UserDemographics model
+ and then removes all duplicate entries except for the most recently modified one.
+ """
+ demographics_model = apps.get_model('demographics', 'UserDemographics')
+ # Retrieve a list of all users that have more than one entry.
+ duplicate_users = (
+ demographics_model.objects.values(
+ 'user'
+ ).annotate(models.Count('id')).values('user').order_by().filter(id__count__gt=1)
+ )
+ # Get a QuerySet of all the UserDemographics instances for the duplicates
+ # sorted by user and modified in descending order.
+ user_demographic_dupes = demographics_model.objects.filter(user__in=duplicate_users).order_by('user', '-modified')
+
+ # Go through the QuerySet and only keep the most recent instance.
+ existing_user_ids = set()
+ for demographic in user_demographic_dupes:
+ if demographic.user_id in existing_user_ids:
+ log.info('UserDemographics {user} -- {modified}'.format(
+ user=demographic.user_id, modified=demographic.modified
+ ))
+ demographic.delete()
+ else:
+ log.info('UserDemographics Duplicate User Delete {user} -- {modified}'.format(
+ user=demographic.user_id, modified=demographic.modified
+ ))
+ existing_user_ids.add(demographic.user_id)
+
+
+class Migration(migrations.Migration):
+ dependencies = [
+ migrations.swappable_dependency(settings.AUTH_USER_MODEL),
+ ('demographics', '0001_initial'),
+ ]
+
+ operations = [
+ migrations.RunPython(_clean_duplicate_entries, migrations.RunPython.noop),
+ ]
diff --git a/openedx/core/djangoapps/demographics/migrations/0003_auto_20200827_1949.py b/openedx/core/djangoapps/demographics/migrations/0003_auto_20200827_1949.py
new file mode 100644
index 00000000000..d31c24841ea
--- /dev/null
+++ b/openedx/core/djangoapps/demographics/migrations/0003_auto_20200827_1949.py
@@ -0,0 +1,20 @@
+# Generated by Django 2.2.15 on 2020-08-27 19:49
+
+from django.conf import settings
+from django.db import migrations, models
+import django.db.models.deletion
+
+
+class Migration(migrations.Migration):
+
+ dependencies = [
+ ('demographics', '0002_clean_duplicate_entries'),
+ ]
+
+ operations = [
+ migrations.AlterField(
+ model_name='userdemographics',
+ name='user',
+ field=models.OneToOneField(on_delete=django.db.models.deletion.CASCADE, to=settings.AUTH_USER_MODEL),
+ ),
+ ]
diff --git a/openedx/core/djangoapps/demographics/models.py b/openedx/core/djangoapps/demographics/models.py
index be8251c80c6..f576b77a632 100644
--- a/openedx/core/djangoapps/demographics/models.py
+++ b/openedx/core/djangoapps/demographics/models.py
@@ -11,7 +11,7 @@ class UserDemographics(TimeStampedModel):
A Users Demographics platform related data in support of the Demographics
IDA and features
"""
- user = models.ForeignKey(User, on_delete=models.CASCADE)
+ user = models.OneToOneField(User, on_delete=models.CASCADE)
show_call_to_action = models.BooleanField(default=True)
history = HistoricalRecords(app='demographics')
--
GitLab