SUPER column support

[rectalogic]

Support for Redshift SUPER column types would be useful.
https://docs.aws.amazon.com/redshift/latest/dg/super-overview.html

Currently we are using something like this (uses SUPER with redshift dialect, and text with other dialects)

import json
import sqlalchemy as sa
import sqlalchemy.ext.compiler
from sqlalchemy import types


class _JSONText(types.Text):
    def bind_expression(self, bindvalue):
        return sa.func.json_parse(bindvalue)


class SUPER(types.TypeDecorator):
    impl = types.Text

    def load_dialect_impl(self, dialect):
        if dialect.name == "redshift":
            return _JSONText()
        return super().load_dialect_impl(dialect)

    def process_bind_param(self, value, dialect):
        if not isinstance(value, str):
            return json.dumps(value)  # should make json.JSONEncoder class configurable here
        return value


@sqlalchemy.ext.compiler.compiles(SUPER, "redshift")
def compile_super_redshift(type_, compiler, **kw):
    return "SUPER"


@sa.event.listens_for(sa.Table, "column_reflect")
def setup_super_type(inspector, table, column_info):
    if isinstance(column_info["type"], types.NullType):
        column_info["type"] = SUPER()

[raajpackt]

How to import the SUPER type such that i could pass it as an argument in the dtype field while using pandas.DataFrame.to_sql ?

[jklukas]

Addressed in #235