SQLAlchemy compatibility

[findepi]

The client should be compatible with SQLAlchemy and there should be tests for this.

[esemeniuc]

Please add some examples or documentation for SQLAlchemy :)

[esemeniuc]

This seems relevant: https://docs.sqlalchemy.org/en/13/core/connections.html#registering-new-dialects

[andreas-eberle]

It looks like @dungdm93 is working on something like this: https://github.com/dungdm93/sqlalchemy-trino

[dungdm93]

Thank you @andreas-eberle. My project still in early stage, any contributions are welcome.

[findepi]

@dungdm93 this is great!
do you have plans to contribute SQLAlchemy support to the main repo here?
cc @joshthoward

[dungdm93]

@findepi I think it's better to keep it in a separated repo like other dialects.
However if possible, I also would like to transfer ownership of that repo to Trino community.

[joshthoward]

@findepi, @dungdm93 IMO we should keep the SQL Alchemy Dialect w/ the DB API implementation. In the link provided tools like pyhive or pydruid do this. I've also noticed that there are a few differences in what SQL Alchemy expects from a DB API implementation vs. what we have (hence the need for bundling them in a single repo).

[shawnzhu]

I had been using the pyHive[presto] as the DB-API and SQLAlchemy driver for presto for a while, it works very well, for instance:

import pandas as pd

sqlalchemy_presto_connection_url = "presto://username:password@presto-coordinator:8080/hive/schema-name"

df = pd.read_sql_query("SHOW tables", sqlalchemy_presto_connection_url)

I'm also interested in this feature since it will make my workflow much better.

[prakharcode]

Hi, what's the movement on this @findepi @dungdm93 ?

[findepi]

@prakharcode there is a PR
@joshthoward knows the current state better

[dungdm93]

Hi @prakharcode, I start moving my code to this repo. First one is just a patch to make cursor.description eager-loading #80