Use LazyTbl without an engine?

[randy3k]

Could we use LazyTbl without specifying an engine? I just want to use siuba to generate SQL query, similar to what pypika does.

[machow]

Hey--does this approach (using a mock engine) work?

from siuba.sql.utils import mock_sqlalchemy_engine
from siuba.sql import LazyTbl as _LazyTbl
cars_sql = _LazyTbl(
        mock_sqlalchemy_engine("postgresql"),
        "cars",
        ["cyl", "mpg", "hp"]
        )

# example showing the query
from siuba import _, mutate, show_query

q = cars_sql >> mutate(avg = _.hp - _.hp.mean()) >> show_query()

outputs:

SELECT cars.cyl, cars.mpg, cars.hp, cars.hp - avg(cars.hp) OVER () AS avg
FROM cars

Currently, show_query just prints the SQL and returns a LazyTbl, but I can add the equivalent of dbplyr's db_sql_render (or add an argument to show_query to return the generated string)

[susannvorberg]

I would also be interested in getting the SQL query and vote for an argument to show_query() to return the generated string.