What is this library ?

This library allows to make a SQL script (or possibly any script) addressable. Once preprocessed or parsed, you can get any chunk of it (which could be a single query for instance) with its line number, key, or index of the chunk.

Example & How to use:

CREATE TEMPORARY VIEW fibonacci (a, index_) AS
WITH RECURSIVE inner_fibo(b, c, index_) AS (
    SELECT 0, 1, 0
    UNION
    SELECT c, c + b, index_ + 1
    FROM inner_fibo
)
SELECT b, index_ FROM inner_fibo;

SELECT * FROM fibonacci LIMIT 1 OFFSET 3;
SELECT * FROM fibonacci LIMIT 1 OFFSET 4;
SELECT * FROM fibonacci LIMIT 1 OFFSET 5;
SELECT * FROM fibonacci LIMIT 1 OFFSET 10;


CREATE TEMPORARY TABLE test_table (a integer) STRICT;

 -- % key

raw_script.sql :

 -- % fibonacci_view_creation
CREATE TEMPORARY VIEW fibonacci (a, index_) AS
WITH RECURSIVE inner_fibo(b, c, index_) AS (
    SELECT 0, 1, 0
    UNION
    SELECT c, c + b, index_ + 1
    FROM inner_fibo
)
SELECT b, index_ FROM inner_fibo;

 -- % contiguous_view_tests
SELECT * FROM fibonacci LIMIT 1 OFFSET 3;
SELECT * FROM fibonacci LIMIT 1 OFFSET 4;
SELECT * FROM fibonacci LIMIT 1 OFFSET 5;
 -- % testing_10nth_value_of_view
SELECT * FROM fibonacci LIMIT 1 OFFSET 10;


CREATE TEMPORARY TABLE test_table (a integer) STRICT;

Then we preprocess the script :

import addressable_sql_queries
with (open("raw_script.sql") as input_file,
    open("preprocessed_script.py", "w") as output_file):
    output_file.write(
        addressable_sql_queries.preprocess_to_str(input_file.read()))

This step can also be done with :

python -m addressable_sql_queries path/to/raw_script.sql -o path/to/preprocessed_script.py

As all keys in the example above are not only valid but usable python ids, you can then access chunks of this script like this :

from preprocessed_script import *
print(contiguous_view_tests)

Which will outputs :

 -- % contiguous_view_tests
SELECT * FROM fibonacci LIMIT 1 OFFSET 3;
SELECT * FROM fibonacci LIMIT 1 OFFSET 4;
SELECT * FROM fibonacci LIMIT 1 OFFSET 5;

Even when keys are invalid ids, you can access the chunks of the script by these means :

from preprocessed_script import queries as sql_queries
print(sql_queries[0]) # This will print the chunk before the first delimiter, which is nothing in this case : "\n".
print(sql_queries[None]) # Another way to do the same thing. None is always the first key.
print(sql_queries["fibonacci_view_creation"])
"""This will print :
 -- % fibonacci_view_creation
CREATE TEMPORARY VIEW fibonacci (a, index_) AS
WITH RECURSIVE inner_fibo(b, c, index_) AS (
    SELECT 0, 1, 0
    UNION
    SELECT c, c + b, index_ + 1
    FROM inner_fibo
)
SELECT b, index_ FROM inner_fibo;


"""
print(sql_queries[1]) # This will print the same thing as above (the second chunk).
print(sql_queries.having_line(4)) # This will print the chunk containing the line 4 (counting from 1), and thus does the same thing as above.

Parser mode example :

Alternatively, parser mode can be used :

from addressable_sql_queries import Query
q = Query.fromFile("raw_script.sql")
print(q[3])
"""This will print :
 -- % testing_10nth_value_of_view
SELECT * FROM fibonacci LIMIT 1 OFFSET 10;


CREATE TEMPORARY TABLE test_table (a integer) STRICT;
"""

Query object can also be built directly from a string.

Note : in the case where the first line is a delimiter, thus making the first and second chunk shares the first line (starting at 1), the having_line function will return the second, non empty chunk.