Skip to content

apply_ddl

Apply every schema and table definition to TimescaleDB.

Files under stock_brokers/instruments/sql/ddl are executed in filename order, so the numeric prefix decides what is created first - schemas before the tables that live in them. Every statement is written to be safe to run again, which makes this both the way to create the tables and the way to pick up a later change to one.

python -m stock_brokers.instruments.sql.apply_ddl

Functions:

Name Description
apply_all

Execute every DDL file in filename order.
main

Apply the DDL and report how many files were run.

apply_all 

apply_all(directory=DDL_DIRECTORY)

Execute every DDL file in filename order.

The whole run is one transaction: a broken definition rolls the batch back rather than leaving the database half migrated.

  • directory is the folder holding the .sql files.
Source code in stock_brokers/instruments/sql/apply_ddl.py 
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
def apply_all(directory=DDL_DIRECTORY):
    """
    Execute every DDL file in filename order.

    The whole run is one transaction: a broken definition rolls the batch back rather than
    leaving the database half migrated.

    - `directory` is the folder holding the `.sql` files.
    """
    if not directory.is_dir():
        raise FileNotFoundError(f"No DDL directory at {directory}")

    paths = sorted(directory.glob("*.sql"))
    with get_postgres_engine().begin() as connection:
        for path in paths:
            connection.execute(text(path.read_text()))
            print(f"  applied {path.name}")
    return len(paths)

main 

main()

Apply the DDL and report how many files were run.

Source code in stock_brokers/instruments/sql/apply_ddl.py 
39
40
41
42
def main():
    """Apply the DDL and report how many files were run."""
    print(f"Applying DDL from {DDL_DIRECTORY}\n")
    print(f"\nApplied {apply_all()} DDL file(s).")