Unable to convert a table to hypertable due to unique index error

Question

I am creating a table using Flask SQL Alchemy as follows:

class Price(db.Model):
    __tablename__ = "prices"
    id = db.Column(db.Integer, primary_key=True)
    country_code = db.Column(db.String(2), nullable=False)
    value = db.Column(db.Float(precision=2), nullable=False)
    start = db.Column(db.DateTime(timezone=True), nullable=False)
    end = db.Column(db.DateTime(timezone=True), nullable=False)
    price_type = db.Column(db.String(32), nullable=False)

and then I want to create the table and convert it into hypertable by doing:

def register_timescale():
    result = db.session.execute(
        "select exists(select * from timescaledb_information.hypertable " +
        " where table_name='prices')")
    is_hypertable = [row[0] for row in result]
    if is_hypertable[0] is False:
        db.session.execute(
            "SELECT create_hypertable('prices', 'start', chunk_time_interval=>INTERVAL '1 day')")
        db.session.commit()


def create_app(app_config):
    app = Flask(__name__)

    app.config.from_object(app_config)

    register_extensions(app)
    register_blueprints(app)
    with app.app_context():
        db.create_all()
        register_timescale()
    return app

I want to create a hypertable of the table prices with the start column being time_column_name as described at create_hypertable and the chunk_time_interval should be 1 day. Meaning, one chunk must have data of 1 day

Example:

One chunk must contain data where start value is 2020-04-19 00:00:00+02 and end value is 2020-04-20 00:00:00+02 (from row 1 to 24 in the below image) and then the second chunk must contain value from 2020-04-20 00:00:00+02 till 2020-04-21 00:00:00+02 (from row 25 to 48 in the below image) and so on.

When I do this, I am getting the following error:

Exception has occurred: DatabaseError (psycopg2.DatabaseError) cannot create a unique index without the column "start" (used in partitioning)

[SQL: SELECT create_hypertable('prices', 'start', chunk_time_interval=>INTERVAL '1 day')] (Background on this error at: http://sqlalche.me/e/4xp6)

Then, I tried changing the register_timescale() function as follows:

def register_timescale():
    result = db.session.execute(
        "select exists(select * from timescaledb_information.hypertable " +
        " where table_name='prices')")
    is_hypertable = [row[0] for row in result]
    if is_hypertable[0] is False:
        db.session.execute(
            "ALTER DATABASE mydb SET timezone TO 'Europe/Berlin'")
        db.session.execute("SELECT pg_reload_conf()")
        db.session.execute("ALTER TABLE prices ADD PRIMARY KEY (id, start)")
        db.session.execute(
            "SELECT create_hypertable('prices', (id, start), chunk_time_interval=>INTERVAL '1 day')")
        db.session.commit()

And this gives me the following error:

Exception has occurred: ProgrammingError (psycopg2.errors.InvalidTableDefinition) multiple primary keys for table "prices" are not allowed

[SQL: ALTER TABLE prices ADD PRIMARY KEY (id, start)] (Background on this error at: http://sqlalche.me/e/f405)

My data looks as follows

Can someone please point out the mistake I am doing?

Thanks

Answer 1

TimescaleDB requires that unique and primary keys include the time dimension column. While only one column can be defined as the time dimension. You can read about limitations in create_hypertable doc.

Since Flask SQL Alchemy requires a primary key, you can define a compound primary key on the time dimension column, which is start in your case, and id. Based on this doc I believe it will be:

class Price(db.Model):
    __tablename__ = "prices"
    id = db.Column(db.Integer, primary_key=True)
    country_code = db.Column(db.String(2), nullable=False)
    value = db.Column(db.Float(precision=2), nullable=False)
    start = db.Column(db.DateTime(timezone=True), primary_key=True)
    end = db.Column(db.DateTime(timezone=True), nullable=False)
    price_type = db.Column(db.String(32), nullable=False)

Then you should be able to execute create_hypertable statement successfully. It should include only column start as the time dimension, i.e.:

SELECT create_hypertable('prices', 'start', chunk_time_interval=>INTERVAL '1 day')

Note that create_hypertable will also create the default index on start. Depending on your application queries you might not need this index and the primary key might be enough. In such case you can drop the index on start either by setting option create_default_indexes to false during the call to create hypertable or by executing DROP INDEX statement afterwards. The best practice section contains more information.