Discussion: `SqliteTask`

I am currently working on a simple implementation of a SqliteTask for interacting with sqlite3 databases.

What I’d like to support:

• executing single statements and returning the values of fetchall() (or some prespecified number of return rows)
• executing full sql scripts (either as a string or read from a file); note that if you execute a script only, there is nothing to return

My current sketch design idea:

class SqliteTask(Task):
    def __init__(self, statement: str = None, 
                       script: str = None, db: str = ":memory:", 
                       return_rows: int = 0, **kwargs):
        ...

Moreover, if you provide both script and statement, the script will be executed first and the statement last, in case it’s a SELECT statement to query what just happened in the script.

I’m opening this issue for discussion in case this is overloading the task with too many parameters / options / possibilities. I want to make sure it’s still clear as a single Task.

cc: @jlowin @joshmeek

[cicdw]

Yea; I don’t think that, in implementing this sqlite functionality, I should use any Prefect conditionals – if we agree the goals are distinct enough, I’ll just make two different tasks that the user has to choose from.

[jlowin]

+1

I’d almost always vote in favor of smaller tasks that can be combined into complex flows than larger tasks that do conditional logic internally.