Test integrating TimescaleDB into Django. It requires a PostgreSQL client and Pipenv to be installed. To start and setup the database.
docker run -d --rm --name timescaledb -p 127.0.0.1:5432:5432 -e POSTGRES_PASSWORD=password timescale/timescaledb:latest-pg12
psql -h 127.0.0.1 -U postgres
--> password
--> CREATE DATABASE sensors;
--> \q
pipenv install
pipenv shell
./manage.py migrate
./manage.py loaddata test_data.json
To start the dev server run ./manage.py runserver
In order to test the smearing of timestamps on collisions, in another terminal, run curl -d "value=<float>" 127.0.0.1:8000/go/ several times while replacing <float> with a number. Then check the resulting timestamps in the terminal output or the admin view. The newely inserted value should always have a timestamp 1 microsecond greater than the last one (timestamp is hard coded into the view).
If you only have Python 3.7 installed, change the definition in Pipfile from python_version = "3.8" to python_version = "3.7"
On Windows run py manage.py to execute the manage.py commands.
If the following error is thrown during a migration, check and update the column name: django.db.utils.ProgrammingError: column "peripheral_component" specified in option timescaledb.compress_segmentby does not exist. Connect to the database server with the above psql command, connect to the database with \c <database_name>, list the tables with \dt and show the table schema of the table to be converterd to a hypertable with \d <table_name>. Sometimes foreign keys have _id appended to their column name. In this example the variables are
- database_name: sensors
- table_name: farms_sensorreading