diff --git a/Makefile b/Makefile
index 310c603..27cb0fb 100644
--- a/Makefile
+++ b/Makefile
@@ -8,6 +8,8 @@ collectstatic: run.jarbas
 
 seed: run.jarbas
 	docker-compose run --rm jarbas python manage.py loaddatasets
-	docker-compose run --rm jarbas python manage.py loadsuppliers
+	docker-compose run --rm jarbas python manage.py reimbursements /tmp/serenata-data/reimbursements.xz
+	docker-compose run --rm jarbas python manage.py companies /tmp/serenata-data/2016-09-03-companies.xz
+	docker-compose run --rm jarbas python manage.py irregularities /tmp/serenata-data/irregularities.xz
 
 run.devel: collectstatic
diff --git a/README.md b/README.md
index 1b85a1c..8b90e79 100644
--- a/README.md
+++ b/README.md
@@ -132,42 +132,45 @@ If you have [Docker](https://docs.docker.com/engine/installation/) (with [Docker
 
 ```console
 $ docker-compose up -d --build
-$	docker-compose run --rm jarbas python manage.py migrate
-$	docker-compose run --rm jarbas python manage.py ceapdatasets
+$ docker-compose run --rm jarbas python manage.py migrate
+$ docker-compose run --rm jarbas python manage.py ceapdatasets
 ```
 
-
 You can access it at [`localhost:80`](http://localhost:80/). However your database starts empty and you still have to collect your static files:
 
 ```console
 $ docker-compose run --rm jarbas python manage.py collectstatic --no-input
 $ docker-compose run --rm jarbas python manage.py loaddatasets
-$ docker-compose run --rm jarbas python manage.py reimbursements
-$ docker-compose run --rm jarbas python manage.py loadsupliers
-$ python manage.py irregularities <path to irregularities.xz file>
+$ docker-compose run --rm jarbas python manage.py reimbursements <path to reimbursements.xz>
+$ docker-compose run --rm jarbas python manage.py irregularities <path to irregularities.xz file>
+$ docker-compose run --rm jarbas python manage.py companies <path to companies.xz>
 ```
 
-There are some cleaver shortcuts in the `Makefile` if you like it.
+You can get the datasets running [Rosie](https://github.com/datasciencebr/rosie) or directly with the [toolbox](https://github.com/datasciencebr/rosie).
+
+Also there are some cleaver shortcuts in the `Makefile` if you like it. 
 
 ### Local install
 
 #### Requirements
 
-The app is based in [Python 3.5](http://python.org) and [Node.js 6](http://nodejs.org). Once you have `pip` and `npm` available, install the dependencies:
+Jarbas requires [Python 3.5](http://python.org), [Node.js 6](http://nodejs.org). and [PostgreSQL 9.4+](https://www.postgresql.org).
+
+Once you have `pip` and `npm` available install the dependencies:
 
 ```console
-npm i
+npm install
 python -m pip install -r requirements.txt
 ```
 
-Minor details on requirements:
+##### Python's `lzma` module
+
+In some Linux distros `lzma` is not installed by default. You can check whether you have it or not with `$ python -m lzma`. In Debian based systems you can fix that with `$ apt-get install liblzma-dev` or in macOS with `$ brew install xz` — but you mihght have to re-compile your Python.
 
-* **`lzma`**: In some Linux distros `lzma` is not installed by default. You can check whether you have it or not with `$ python -m lzma`. In Debian based systems you can fix that with `$ apt-get install liblzma-dev` but you mihght have to re-compile your Python. Some macOS Users might have the same problem. To check if you have `lzma` you can use `$ python -m lmza`. To fix it you need to install `lzma` using `$ brew install xz` and after that you need to recompile Python, and an way to do it is through `$ brew upgrade --cleanup python`.
-* **`psycopg2`**: The `requirements.txt` file is prepared to use [PostgresSQL](https://www.postgresql.org) and `psycopg2` might fail if you don't have Postgres installed locally.
 
 #### Settings
 
-Copy `contrib/.env.sample` as `.env` in the project's root folder and adjust your settings. These are the main environment settings:
+Copy `contrib/.env.sample` as `.env` in the project's root folder and adjust your settings. These are the main variables:
 
 ##### Django settings
 
@@ -187,8 +190,6 @@ Copy `contrib/.env.sample` as `.env` in the project's root folder and adjust you
 * `AMAZON_S3_BUCKET` (_str_) Name of the Amazon S3 bucket to look for datasets (e.g. `serenata-de-amor-data`)
 * `AMAZON_S3_REGION` (_str_) Region of the Amazon S3 (e.g. `s3-sa-east-1`)
 * `AMAZON_S3_DATASET_DATE` (_str_) Datasets file name prefix of CEAP datasets from Serenata de Amor (e.g. `2016-08-08` for `2016-08-08-current-year.xz`)
-* `AMAZON_S3_REIMBURSEMENTS_DATE` (_str_) Reumbursements dataset file name date prefix (e.g. `2016-12-06` for `2016-12-06-reimbursements.xz`)
-* `AMAZON_S3_COMPANIES_DATE` (_str_) Suppliers (companies) datasets file name date prefix (e.g. `2016-08-08` for `2016-08-08-companies.xz`)
 * `AMAZON_S3_CEAPTRANSLATION_DATE` (_str_) File name prefix for dataset guide (e.g. `2016-08-08` for `2016-08-08-ceap-datasets.md`)
 
 ##### Google settings
@@ -210,19 +211,13 @@ Now you can load the data from our datasets and get some other data as static fi
 
 ```
 $ python manage.py loaddatasets
-$ python manage.py loadsuppliers
-$ python manage.py reimbursements
+$ python manage.py reimbursements <path to reimbursements.xz>
+$ python manage.py irregularities <path to irregularities.xz file>
+$ python manage.py companies <path to companies.xz>
 $ python manage.py ceapdatasets
 ```
 
-Use `python manage.py loaddatasets --help` and `python manage.py loadsuppliers --help` to check options on limiting the number of documents to be loaded from the datasets.
-
-If [Rosie](https://github.com/datasciencebr/rosie) was kind enough to give you
-a `irregularities.xz`, you can load it with:
-
-```
-$ python manage.py irregularities <path to irregularities.xz file>
-```
+You can get the datasets running [Rosie](https://github.com/datasciencebr/rosie) or directly with the [toolbox](https://github.com/datasciencebr/rosie).
 
 #### Generate static files
 
diff --git a/contrib/.env.sample b/contrib/.env.sample
index 44e84a0..daf4ed6 100644
--- a/contrib/.env.sample
+++ b/contrib/.env.sample
@@ -9,7 +9,5 @@ AMAZON_S3_REGION=s3-sa-east-1
 
 AMAZON_S3_CEAPTRANSLATION_DATE=2016-08-08
 AMAZON_S3_DATASET_DATE=2016-08-08
-AMAZON_S3_REIMBURSEMENTS_DATE=2016-12-06
-AMAZON_S3_SUPPLIERS_DATE=2016-09-03
 
 GOOGLE_STREET_VIEW_API_KEY=my-google-places-api-key
diff --git a/jarbas/core/management/commands/__init__.py b/jarbas/core/management/commands/__init__.py
index c268740..0177a4a 100644
--- a/jarbas/core/management/commands/__init__.py
+++ b/jarbas/core/management/commands/__init__.py
@@ -10,53 +10,13 @@
 
 class LoadCommand(BaseCommand):
 
-    def add_arguments(self, parser):
-        parser.add_argument(
-            '--source', '-s', dest='source', default=None,
-            help='Data directory of Serenata de Amor (dataset source)'
-        )
-        parser.add_argument(
-            '--drop-all', '-d', dest='drop', action='store_true',
-            help='Drop all existing records before loading the datasets'
-        )
-        parser.add_argument(
-            '--dataset-version', dest='dataset_version', default=None,
-            help='Dataset file version (usualy a YYYY-MM-DD date)'
-        )
-
-    def get_dataset(self, name):
-        if self.source:
-            return self.load_local(self.source, name)
-        return self.load_remote(name)
-
-    def load_remote(self, name):
-        """Load a document from Amazon S3"""
-        url = self.get_url(name)
-        print("Loading " + url)
-        with NamedTemporaryFile(delete=False) as tmp:
-            urlretrieve(url, filename=tmp.name)
-            return tmp.name
-
-    def load_local(self, source, name):
-        """Load documents from local source"""
-        path = self.get_path(source, name)
-
-        if not os.path.exists(path):
-            print(path + " not found")
-            return None
-
-        print("Loading " + path)
-        return path
-
-    def get_url(self, suffix):
-        return 'https://{region}.amazonaws.com/{bucket}/{file_name}'.format(
-            region=settings.AMAZON_S3_REGION,
-            bucket=settings.AMAZON_S3_BUCKET,
-            file_name=self.get_file_name(suffix)
-        )
-
-    def get_path(self, source, name):
-        return os.path.join(source, self.get_file_name(name))
+    def add_arguments(self, parser, add_drop_all=True):
+        parser.add_argument('dataset', help='Path to the .xz dataset')
+        if add_drop_all:
+            parser.add_argument(
+                '--drop-all', '-d', dest='drop', action='store_true',
+                help='Drop all existing records before loading the datasets'
+            )
 
     @staticmethod
     def to_number(value, cast=None):
@@ -91,12 +51,6 @@ def to_date(text):
         except ValueError:
             return None
 
-    def get_file_name(self, name):
-        if not self.date:
-            settings_name = 'AMAZON_S3_{}_DATE'.format(name.upper())
-            self.date = getattr(settings, settings_name)
-        return '{date}-{name}.xz'.format(date=self.date, name=name)
-
     def drop_all(self, model):
         if model.objects.count() != 0:
             msg = 'Deleting all existing records from {} model'
@@ -115,3 +69,61 @@ def print_count(self, model, **kwargs):
     @staticmethod
     def get_model_name(model):
         return model._meta.label.split('.')[-1]
+
+
+class OldLoadCommand(LoadCommand):
+
+    def add_arguments(self, parser):
+        parser.add_argument(
+            '--drop-all', '-d', dest='drop', action='store_true',
+            help='Drop all existing records before loading the datasets'
+        )
+        parser.add_argument(
+            '--source', '-s', dest='source', default=None,
+            help='Data directory of Serenata de Amor (dataset source)'
+        )
+        parser.add_argument(
+            '--dataset-version', dest='dataset_version', default=None,
+            help='Dataset file version (usualy a YYYY-MM-DD date)'
+        )
+
+    def get_dataset(self, name):
+        if self.source:
+            return self.load_local(self.source, name)
+        return self.load_remote(name)
+
+    def load_remote(self, name):
+        """Load a document from Amazon S3"""
+        url = self.get_url(name)
+        print("Loading " + url)
+        with NamedTemporaryFile(delete=False) as tmp:
+            urlretrieve(url, filename=tmp.name)
+            return tmp.name
+
+    def load_local(self, source, name):
+        """Load documents from local source"""
+        path = self.get_path(source, name)
+
+        if not os.path.exists(path):
+            print(path + " not found")
+            return None
+
+        print("Loading " + path)
+        return path
+
+    def get_url(self, suffix):
+        return 'https://{region}.amazonaws.com/{bucket}/{file_name}'.format(
+            region=settings.AMAZON_S3_REGION,
+            bucket=settings.AMAZON_S3_BUCKET,
+            file_name=self.get_file_name(suffix)
+        )
+
+    def get_path(self, source, name):
+        return os.path.join(source, self.get_file_name(name))
+
+    def get_file_name(self, name):
+        if not self.date:
+            settings_name = 'AMAZON_S3_{}_DATE'.format(name.upper())
+            self.date = getattr(settings, settings_name)
+        return '{date}-{name}.xz'.format(date=self.date, name=name)
+
diff --git a/jarbas/core/management/commands/loadsuppliers.py b/jarbas/core/management/commands/companies.py
similarity index 90%
rename from jarbas/core/management/commands/loadsuppliers.py
rename to jarbas/core/management/commands/companies.py
index 2300129..6956f9b 100644
--- a/jarbas/core/management/commands/loadsuppliers.py
+++ b/jarbas/core/management/commands/companies.py
@@ -12,10 +12,8 @@ class Command(LoadCommand):
     help = 'Load Serenata de Amor supplier dataset into the database'
 
     def handle(self, *args, **options):
-        self.date = options.get('dataset_version')
-        self.source = options.get('source')
+        self.path = options['dataset']
         self.count = self.print_count(Supplier)
-        print('self.cont =', self.count)
         print('Starting with {:,} suppliers'.format(self.count))
 
         if options.get('drop', False):
@@ -23,16 +21,16 @@ def handle(self, *args, **options):
             self.drop_all(Activity)
             self.count = 0
 
-        self.save_suppliers(self.get_dataset('companies'))
+        self.save_suppliers()
 
-    def save_suppliers(self, dataset):
+    def save_suppliers(self):
         """
         Receives path to the dataset file and create a Supplier object for
         each row of each file. It creates the related activity when needed.
         """
         skip = ('main_activity', 'secondary_activty')
         keys = list(f.name for f in Supplier._meta.fields if f not in skip)
-        with lzma.open(dataset, mode='rt') as file_handler:
+        with lzma.open(self.path, mode='rt') as file_handler:
             for row in csv.DictReader(file_handler):
                 main, secondary = self.save_activities(row)
 
diff --git a/jarbas/core/management/commands/irregularities.py b/jarbas/core/management/commands/irregularities.py
index f1a4f9c..d1dab84 100644
--- a/jarbas/core/management/commands/irregularities.py
+++ b/jarbas/core/management/commands/irregularities.py
@@ -12,14 +12,10 @@ class Command(LoadCommand):
     filter_keys = ('applicant_id', 'document_id', 'year')
 
     def add_arguments(self, parser):
-        parser.add_argument(
-            '--irregularities', '-i', dest='irregularities_path',
-            default='irregularities.xz',
-            help='Path to the irregularities.xz dataset'
-        )
+        super().add_arguments(parser, add_drop_all=False)
 
     def handle(self, *args, **options):
-        self.path = options.get('irregularities_path', 'irregularities.xz')
+        self.path = options['dataset']
         if not os.path.exists(self.path):
             raise FileNotFoundError(os.path.abspath(self.path))
 
diff --git a/jarbas/core/management/commands/loaddatasets.py b/jarbas/core/management/commands/loaddatasets.py
index c9cb227..054a44d 100644
--- a/jarbas/core/management/commands/loaddatasets.py
+++ b/jarbas/core/management/commands/loaddatasets.py
@@ -5,11 +5,11 @@
 
 from django.conf import settings
 
-from jarbas.core.management.commands import LoadCommand
+from jarbas.core.management.commands import OldLoadCommand
 from jarbas.core.models import Document
 
 
-class Command(LoadCommand):
+class Command(OldLoadCommand):
     help = 'Load Serenata de Amor datasets into the database'
     suffixes = ('current-year', 'last-year', 'previous-years')
 
diff --git a/jarbas/core/management/commands/reimbursements.py b/jarbas/core/management/commands/reimbursements.py
index e38534f..620fb96 100644
--- a/jarbas/core/management/commands/reimbursements.py
+++ b/jarbas/core/management/commands/reimbursements.py
@@ -16,8 +16,7 @@ def add_arguments(self, parser):
         )
 
     def handle(self, *args, **options):
-        self.date = options.get('dataset_version')
-        self.source = options.get('source')
+        self.path = options['dataset']
         self.count = Reimbursement.objects.count()
         print('Starting with {:,} reimbursements'.format(self.count))
 
@@ -31,8 +30,7 @@ def handle(self, *args, **options):
     @property
     def reimbursements(self):
         """Returns a Generator with a Reimbursement object for each row."""
-        dataset = self.get_dataset('reimbursements')
-        with lzma.open(dataset, mode='rt') as file_handler:
+        with lzma.open(self.path, mode='rt') as file_handler:
             for row in csv.DictReader(file_handler):
                 yield Reimbursement(**self.serialize(row))
 
diff --git a/jarbas/core/tests/test_loadsuppliers_command.py b/jarbas/core/tests/test_companies_command.py
similarity index 65%
rename from jarbas/core/tests/test_loadsuppliers_command.py
rename to jarbas/core/tests/test_companies_command.py
index f6fa4e2..086e18b 100644
--- a/jarbas/core/tests/test_loadsuppliers_command.py
+++ b/jarbas/core/tests/test_companies_command.py
@@ -4,7 +4,7 @@
 
 from django.test import TestCase
 
-from jarbas.core.management.commands.loadsuppliers import Command
+from jarbas.core.management.commands.companies import Command
 from jarbas.core.models import Activity, Supplier
 from jarbas.core.tests import sample_supplier_data
 
@@ -61,11 +61,11 @@ def test_save_activities(self, update_or_create):
         self.assertEqual(1, len(main))
         self.assertEqual(99, len(secondaries))
 
-    @patch('jarbas.core.management.commands.loadsuppliers.lzma')
-    @patch('jarbas.core.management.commands.loadsuppliers.csv.DictReader')
-    @patch('jarbas.core.management.commands.loadsuppliers.Command.save_activities')
-    @patch('jarbas.core.management.commands.loadsuppliers.Command.serialize')
-    @patch('jarbas.core.management.commands.loadsuppliers.Command.print_count')
+    @patch('jarbas.core.management.commands.companies.lzma')
+    @patch('jarbas.core.management.commands.companies.csv.DictReader')
+    @patch('jarbas.core.management.commands.companies.Command.save_activities')
+    @patch('jarbas.core.management.commands.companies.Command.serialize')
+    @patch('jarbas.core.management.commands.companies.Command.print_count')
     @patch.object(Supplier.objects, 'create')
     def test_save_suppliers(self, create, print_count, serialize, save_activities, rows, lzma):
         self.command.count = 0
@@ -73,7 +73,8 @@ def test_save_suppliers(self, create, print_count, serialize, save_activities, r
         rows.return_value = [sample_supplier_data]
         serialize.return_value = dict(ahoy=42)
         save_activities.return_value = ([3], [14, 15])
-        self.command.save_suppliers([])
+        self.command.path = 'companies.xz'
+        self.command.save_suppliers()
         create.assert_called_with(ahoy=42)
         create.return_value.main_activity.add.assert_called_with(3)
         self.assertEqual(2, create.return_value.secondary_activity.add.call_count)
@@ -81,30 +82,26 @@ def test_save_suppliers(self, create, print_count, serialize, save_activities, r
 
 class TestConventionMethods(TestCommand):
 
-    @patch('jarbas.core.management.commands.loadsuppliers.print')
-    @patch('jarbas.core.management.commands.loadsuppliers.LoadCommand.load_remote')
-    @patch('jarbas.core.management.commands.loadsuppliers.LoadCommand.drop_all')
-    @patch('jarbas.core.management.commands.loadsuppliers.Command.save_suppliers')
-    @patch('jarbas.core.management.commands.loadsuppliers.Command.print_count')
-    def test_handler_without_options(self, print_count, save_suppliers, drop_all, load_remote, print_):
+    @patch('jarbas.core.management.commands.companies.print')
+    @patch('jarbas.core.management.commands.companies.LoadCommand.drop_all')
+    @patch('jarbas.core.management.commands.companies.Command.save_suppliers')
+    @patch('jarbas.core.management.commands.companies.Command.print_count')
+    def test_handler_without_options(self, print_count, save_suppliers, drop_all, print_):
         print_count.return_value = 0
-        self.command.handle(source=None)
+        self.command.handle(dataset='companies.xz')
         print_.assert_called_with('Starting with 0 suppliers')
-        self.assertEqual(1, load_remote.call_count)
         self.assertEqual(1, save_suppliers.call_count)
         self.assertEqual(1, print_count.call_count)
+        self.assertEqual('companies.xz', self.command.path)
         drop_all.assert_not_called()
 
-    @patch('jarbas.core.management.commands.loadsuppliers.print')
-    @patch('jarbas.core.management.commands.loadsuppliers.LoadCommand.load_local')
-    @patch('jarbas.core.management.commands.loadsuppliers.Command.drop_all')
-    @patch('jarbas.core.management.commands.loadsuppliers.Command.save_suppliers')
-    @patch('jarbas.core.management.commands.loadsuppliers.Command.print_count')
-    def test_handler_with_options(self, print_count, save_suppliers, drop_all, load_local, print_):
+    @patch('jarbas.core.management.commands.companies.print')
+    @patch('jarbas.core.management.commands.companies.Command.drop_all')
+    @patch('jarbas.core.management.commands.companies.Command.save_suppliers')
+    @patch('jarbas.core.management.commands.companies.Command.print_count')
+    def test_handler_with_options(self, print_count, save_suppliers, drop_all, print_):
         print_count.return_value = 0
-        self.command.handle(source='ahoy', drop=True, dataset_version='1')
+        self.command.handle(dataset='companies.xz', drop=True)
         print_.assert_called_with('Starting with 0 suppliers')
         self.assertEqual(2, drop_all.call_count)
-        self.assertEqual(1, load_local.call_count)
         self.assertEqual(1, save_suppliers.call_count)
-        self.assertEqual('1', self.command.date)
diff --git a/jarbas/core/tests/test_irregularities_command.py b/jarbas/core/tests/test_irregularities_command.py
index 8385c42..9f2fe2a 100644
--- a/jarbas/core/tests/test_irregularities_command.py
+++ b/jarbas/core/tests/test_irregularities_command.py
@@ -84,29 +84,19 @@ class TestConventionMethods(TestCommand):
     @patch('jarbas.core.management.commands.irregularities.Command.update')
     @patch('jarbas.core.management.commands.irregularities.os.path.exists')
     @patch('jarbas.core.management.commands.irregularities.print')
-    def test_handler_without_options(self, print_, exists, update, irregularities):
-        self.command.handle()
+    def test_handler(self, print_, exists, update, irregularities):
+        self.command.handle(dataset='irregularities.xz')
         update.assert_called_once_with(irregularities)
         print_.assert_called_once_with('0 reimbursements updated.')
         self.assertEqual(self.command.path, 'irregularities.xz')
 
-    @patch('jarbas.core.management.commands.irregularities.Command.irregularities')
-    @patch('jarbas.core.management.commands.irregularities.Command.update')
-    @patch('jarbas.core.management.commands.irregularities.os.path.exists')
-    @patch('jarbas.core.management.commands.irregularities.print')
-    def test_handler_with_options(self, print_, exists, update, irregularities):
-        self.command.handle(irregularities_path='0')
-        update.assert_called_once_with(irregularities)
-        print_.assert_called_once_with('0 reimbursements updated.')
-        self.assertEqual('0', self.command.path)
-
     @patch('jarbas.core.management.commands.irregularities.Command.irregularities')
     @patch('jarbas.core.management.commands.irregularities.Command.update')
     @patch('jarbas.core.management.commands.irregularities.os.path.exists')
     def test_handler_with_non_existing_file(self, exists, update, irregularities):
         exists.return_value = False
         with self.assertRaises(FileNotFoundError):
-            self.command.handle()
+            self.command.handle(dataset='irregularities.xz')
         update.assert_not_called()
 
 
@@ -115,8 +105,7 @@ class TestFileLoader(TestCommand):
     @patch('jarbas.core.management.commands.irregularities.lzma')
     @patch('jarbas.core.management.commands.irregularities.csv.DictReader')
     @patch('jarbas.core.management.commands.irregularities.Command.serialize')
-    @patch('jarbas.core.management.commands.irregularities.Command.get_dataset')
-    def test_irregularities_property(self, get_dataset, serialize, rows, lzma):
+    def test_irregularities_property(self, serialize, rows, lzma):
         lzma.return_value = StringIO()
         rows.return_value = range(42)
         self.command.path = 'irregularities.xz'
diff --git a/jarbas/core/tests/test_load_command.py b/jarbas/core/tests/test_load_command.py
index 4152e88..d2835f4 100644
--- a/jarbas/core/tests/test_load_command.py
+++ b/jarbas/core/tests/test_load_command.py
@@ -13,12 +13,6 @@ class TestStaticMethods(TestCase):
     def setUp(self):
         self.cmd = LoadCommand()
 
-    def test_get_file_name(self):
-        self.cmd.date = None
-        expected = '1970-01-01-companies.xz'
-        with self.settings(AMAZON_S3_COMPANIES_DATE='1970-01-01'):
-            self.assertEqual(expected, self.cmd.get_file_name('companies'))
-
     def test_get_model_name(self):
         self.assertEqual('Activity', self.cmd.get_model_name(Activity))
 
@@ -79,66 +73,14 @@ def test_drop_all(self, mock_print):
         self.assertEqual(0, Activity.objects.count())
 
 
-class TestLocalMethods(TestCase):
-
-    def setUp(self):
-        self.cmd = LoadCommand()
-        self.source = '/whatever/works'
-        self.name = 'companies'
-
-    def test_get_path(self):
-        self.cmd.date = None
-        expected = '/whatever/works/1970-01-01-companies.xz'
-        with self.settings(AMAZON_S3_COMPANIES_DATE='1970-01-01'):
-            result = self.cmd.get_path(self.source, self.name)
-            self.assertEqual(expected, result)
-
-    @patch('jarbas.core.management.commands.print')
-    @patch('jarbas.core.management.commands.os.path.exists')
-    def test_load_local_exists(self, mock_exists, mock_print):
-        self.cmd.date = None
-        mock_exists.return_value = True
-        self.assertIsInstance(self.cmd.load_local(self.source, self.name), str)
-
-    @patch('jarbas.core.management.commands.print')
-    @patch('jarbas.core.management.commands.os.path.exists')
-    def test_load_local_fail(self, mock_exists, mock_print):
-        self.cmd.date = None
-        mock_exists.return_value = False
-        self.assertFalse(self.cmd.load_local(self.source, self.name))
-
-
-class TestRemoteMethods(TestCase):
-
-    def setUp(self):
-        self.cmd = LoadCommand()
-        self.name = 'companies'
-        self.url = 'https://south.amazonaws.com/jarbas/1970-01-01-companies.xz'
-        self.custom_settings = {
-            'AMAZON_S3_COMPANIES_DATE': '1970-01-01',
-            'AMAZON_S3_REGION': 'south',
-            'AMAZON_S3_BUCKET': 'jarbas'
-        }
-
-    def test_get_url(self):
-        self.cmd.date = None
-        with self.settings(**self.custom_settings):
-            result = self.cmd.get_url(self.name)
-            self.assertEqual(self.url, result)
-
-    @patch('jarbas.core.management.commands.print')
-    @patch('jarbas.core.management.commands.urlretrieve')
-    def test_load_remote(self, mock_urlretrieve, mock_print):
-        self.cmd.date = None
-        with self.settings(**self.custom_settings):
-            result = self.cmd.load_remote(self.name)
-            self.assertEqual(self.url, mock_urlretrieve.call_args[0][0])
-            self.assertIsInstance(result, str)
-
-
 class TestAddArguments(TestCase):
 
     def test_add_arguments(self):
         mock = Mock()
         LoadCommand().add_arguments(mock)
-        self.assertEqual(3, mock.add_argument.call_count)
\ No newline at end of file
+        self.assertEqual(2, mock.add_argument.call_count)
+
+    def test_add_arguments_without_drop_all(self):
+        mock = Mock()
+        LoadCommand().add_arguments(mock, add_drop_all=False)
+        self.assertEqual(1, mock.add_argument.call_count)
\ No newline at end of file
diff --git a/jarbas/core/tests/test_loaddatasets_command.py b/jarbas/core/tests/test_loaddatasets_command.py
index ddbb00c..2c7c450 100644
--- a/jarbas/core/tests/test_loaddatasets_command.py
+++ b/jarbas/core/tests/test_loaddatasets_command.py
@@ -88,7 +88,7 @@ def test_get_suffix(self):
         self.assertEqual(self.command.get_suffix(name), 'current-year')
         self.assertEqual(self.command.get_suffix(''), None)
 
-    @patch('jarbas.core.management.commands.loaddatasets.LoadCommand.load_local')
+    @patch('jarbas.core.management.commands.loaddatasets.OldLoadCommand.load_local')
     def test_get_load_local(self, super_load_local):
         list(self.command.load_local('ahoy'))
         expected = (
@@ -98,7 +98,7 @@ def test_get_load_local(self, super_load_local):
         )
         super_load_local.assert_has_calls(expected)
 
-    @patch('jarbas.core.management.commands.loaddatasets.LoadCommand.load_remote')
+    @patch('jarbas.core.management.commands.loaddatasets.OldLoadCommand.load_remote')
     def test_get_load_remote(self, super_load_remote):
         list(self.command.load_remote())
         expected = (
@@ -168,7 +168,7 @@ def test_handler_with_options(self, drop_all, bulk_create_by, documents_from, lo
         bulk_create_by.assert_called_once_with((1, 2, 3), 42)
         self.assertEqual('1', self.command.date)
 
-    @patch('jarbas.core.management.commands.loaddatasets.LoadCommand.add_arguments')
+    @patch('jarbas.core.management.commands.loaddatasets.OldLoadCommand.add_arguments')
     def test_add_arguments(self, super_add_arguments):
         parser = MagicMock()
         self.command.add_arguments(parser)
diff --git a/jarbas/core/tests/test_old_load_command.py b/jarbas/core/tests/test_old_load_command.py
new file mode 100644
index 0000000..bdb330a
--- /dev/null
+++ b/jarbas/core/tests/test_old_load_command.py
@@ -0,0 +1,101 @@
+from unittest.mock import Mock, patch
+
+from django.test import TestCase
+
+from jarbas.core.management.commands import OldLoadCommand
+
+
+class TestFileLoader(TestCase):
+
+    def setUp(self):
+        self.cmd = OldLoadCommand()
+        self.cmd.date = '1970-01-01'
+
+    @patch('jarbas.core.management.commands.OldLoadCommand.load_remote')
+    @patch('jarbas.core.management.commands.print')
+    def test_get_database(self, print_, load_remote):
+        self.cmd.source = None
+        self.cmd.get_dataset('dataset')
+        load_remote.assert_called_once_with('dataset')
+
+    @patch('jarbas.core.management.commands.os.path.exists')
+    @patch('jarbas.core.management.commands.print')
+    def test_get_database_with_source(self, print_, exists):
+        exists.return_value = True
+        self.cmd.source = '/whatever/works'
+        expected = '/whatever/works/1970-01-01-dataset.xz'
+        self.assertEqual(expected, self.cmd.get_dataset('dataset'))
+        print_.assert_called_once_with('Loading ' + expected)
+
+    @patch('jarbas.core.management.commands.os.path.exists')
+    @patch('jarbas.core.management.commands.print')
+    def test_get_database_with_wrong_source(self, print_, exists):
+        exists.return_value = False
+        self.cmd.source = '/whatever/works'
+        expected = '/whatever/works/1970-01-01-dataset.xz'
+        self.assertIsNone(self.cmd.get_dataset('dataset'))
+        print_.assert_called_once_with(expected + ' not found')
+
+class TestLocalMethods(TestCase):
+
+    def setUp(self):
+        self.cmd = OldLoadCommand()
+        self.source = '/whatever/works'
+        self.name = 'dataset'
+
+    def test_get_path(self):
+        self.cmd.date = None
+        expected = '/whatever/works/1970-01-01-dataset.xz'
+        with self.settings(AMAZON_S3_DATASET_DATE='1970-01-01'):
+            result = self.cmd.get_path(self.source, self.name)
+            self.assertEqual(expected, result)
+
+    @patch('jarbas.core.management.commands.print')
+    @patch('jarbas.core.management.commands.os.path.exists')
+    def test_load_local_exists(self, mock_exists, mock_print):
+        self.cmd.date = None
+        mock_exists.return_value = True
+        self.assertIsInstance(self.cmd.load_local(self.source, self.name), str)
+
+    @patch('jarbas.core.management.commands.print')
+    @patch('jarbas.core.management.commands.os.path.exists')
+    def test_load_local_fail(self, mock_exists, mock_print):
+        self.cmd.date = None
+        mock_exists.return_value = False
+        self.assertFalse(self.cmd.load_local(self.source, self.name))
+
+
+class TestRemoteMethods(TestCase):
+
+    def setUp(self):
+        self.cmd = OldLoadCommand()
+        self.name = 'companies'
+        self.url = 'https://south.amazonaws.com/jarbas/1970-01-01-companies.xz'
+        self.custom_settings = {
+            'AMAZON_S3_COMPANIES_DATE': '1970-01-01',
+            'AMAZON_S3_REGION': 'south',
+            'AMAZON_S3_BUCKET': 'jarbas'
+        }
+
+    def test_get_url(self):
+        self.cmd.date = None
+        with self.settings(**self.custom_settings):
+            result = self.cmd.get_url(self.name)
+            self.assertEqual(self.url, result)
+
+    @patch('jarbas.core.management.commands.print')
+    @patch('jarbas.core.management.commands.urlretrieve')
+    def test_load_remote(self, mock_urlretrieve, mock_print):
+        self.cmd.date = None
+        with self.settings(**self.custom_settings):
+            result = self.cmd.load_remote(self.name)
+            self.assertEqual(self.url, mock_urlretrieve.call_args[0][0])
+            self.assertIsInstance(result, str)
+
+
+class TestAddArguments(TestCase):
+
+    def test_add_arguments(self):
+        mock = Mock()
+        OldLoadCommand().add_arguments(mock)
+        self.assertEqual(3, mock.add_argument.call_count)
\ No newline at end of file
diff --git a/jarbas/core/tests/test_reimbursements_command.py b/jarbas/core/tests/test_reimbursements_command.py
index f997125..c898fe5 100644
--- a/jarbas/core/tests/test_reimbursements_command.py
+++ b/jarbas/core/tests/test_reimbursements_command.py
@@ -119,9 +119,10 @@ class TestConventionMethods(TestCommand):
     @patch('jarbas.core.management.commands.reimbursements.Command.print_count')
     def test_handler_without_options(self, print_count, bulk_create_by, reimbursements, print_):
         reimbursements.return_value = (1, 2, 3)
-        self.command.handle(batch_size=42, source=False)
+        self.command.handle(dataset='reimbursements.xz', batch_size=42)
         print_.assert_called_once_with('Starting with 0 reimbursements')
         bulk_create_by.assert_called_once_with(reimbursements, 42)
+        self.assertEqual('reimbursements.xz', self.command.path)
 
     @patch('jarbas.core.management.commands.reimbursements.print')
     @patch('jarbas.core.management.commands.reimbursements.Command.reimbursements')
@@ -129,11 +130,10 @@ def test_handler_without_options(self, print_count, bulk_create_by, reimbursemen
     @patch('jarbas.core.management.commands.reimbursements.Command.drop_all')
     @patch('jarbas.core.management.commands.reimbursements.Command.print_count')
     def test_handler_with_options(self, print_count, drop_all, bulk_create_by, reimbursements, print_):
-        self.command.handle(batch_size=1, source='ahoy', drop=True, dataset_version='1')
+        self.command.handle(dataset='reimbursements.xz', batch_size=1, drop=True)
         print_.assert_called_once_with('Starting with 0 reimbursements')
         drop_all.assert_called_once_with(Reimbursement)
         bulk_create_by.assert_called_once_with(reimbursements, 1)
-        self.assertEqual('1', self.command.date)
 
     @patch('jarbas.core.management.commands.reimbursements.LoadCommand.add_arguments')
     def test_add_arguments(self, super_add_arguments):
@@ -149,9 +149,9 @@ class TestFileLoader(TestCommand):
     @patch('jarbas.core.management.commands.reimbursements.csv.DictReader')
     @patch('jarbas.core.management.commands.reimbursements.Reimbursement')
     @patch('jarbas.core.management.commands.reimbursements.Command.serialize')
-    @patch('jarbas.core.management.commands.reimbursements.Command.get_dataset')
-    def test_reimbursement_property(self, get_dataset, serializer, reimbursement, row, lzma):
+    def test_reimbursement_property(self, serializer, reimbursement, row, lzma):
         lzma.return_value = StringIO()
         row.return_value = dict(ahoy=42)
+        self.command.path = 'reimbursements.xz'
         list(self.command.reimbursements)
         self.assertEqual(1, reimbursement.call_count)
diff --git a/jarbas/settings.py b/jarbas/settings.py
index 5e5d9e8..5a42567 100644
--- a/jarbas/settings.py
+++ b/jarbas/settings.py
@@ -136,10 +136,8 @@
 
 AMAZON_S3_BUCKET = config('AMAZON_S3_BUCKET', default='serenata-de-amor-data')
 AMAZON_S3_REGION = config('AMAZON_S3_REGIN', default='s3-sa-east-1')
+
 AMAZON_S3_DATASET_DATE = config('AMAZON_S3_DATASET_DATE', default='2016-11-19')
-AMAZON_S3_REIMBURSEMENTS_DATE = config('AMAZON_S3_REIMBURSEMENTS_DATE', default='2016-12-06')
-AMAZON_S3_IRREGULARITIES_DATE = config('AMAZON_S3_IRREGULARITIES_DATE', default='2016-12-11')
-AMAZON_S3_COMPANIES_DATE = config('AMAZON_S3_COMPANIES_DATE', default='2016-09-03')
 AMAZON_S3_CEAPTRANSLATION_DATE = config('AMAZON_S3_CEAPTRANSLATION_DATE', default='2016-08-08')
 
 # Django REST Framework