[KED-2141] Impala support for kedro, custom SQLdataset

Description

To use kedro with impala. Is is possible to extend the current SQLDataset to create a connection instead of accepting a string only?

Context

Integration with Impala is useful.

Possible Implementation

I am using the python package impyla. To query something with pandas, I need to first create a connection.

from impala.dbapi import connect

user=user
password=password
host=host

conn = connect(host=host, port=21050, user=user, password=password)
sql = 'select * from trable'
pd.read_sql(sql, conn)

In theory, as long as I can implement the connection creation logic manually, the rest should just match the standard pandasSQLdataset.

[noklam]

Cool, see if I can come up with a PR this week or next week.

On Wed, 10 Mar 2021 at 5:12 PM, Merel Theisen [e-mail] wrote:

@noklam https://github.com/noklam This isn’t a priority for the team at the moment, so we’d be very happy to accept a PR.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/quantumblacklabs/kedro/issues/504#issuecomment-795127374, or unsubscribe https://github.com/notifications/unsubscribe-auth/AELAWL5S5DDXFB6IUHFBJQ3TC4ZYHANCNFSM4Q7NYALQ .

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.