Psycopg2插入与更新使用字典值冲突

Question

I have some large datasets to perform the Postgres upsert function. I am trying to sort out how to dynamically create the columns/values off my dictionary. I am able to insert using the As Is extension using:

sql = "INSERT INTO table (%s) VALUES %s;"
cols = mydict.keys()
vals = [mydict[col] for col in cols]
cursor.execute(sql, (AsIs(', '.join(cols)), tuple(vals),))

This seems to work fine if I am not doing any sort of conflict resolution. or saying ON CONFLICT ( col1) DO NOTHING. But I need to do an UPDATE of the values.

sql = "INSERT INTO table (%s) VALUES %s ON CONFLICT (col1) DO UPDATE SET ....???"

So I am not sure if there is a good method to created that SET col1=val1 type set other than manually creating %(col1)s = %(val1)s variables in the SQL.

Answer 1

Using tools from here:

https://www.psycopg.org/docs/sql.html

An example of something I do:

# div_flds is list of field names
ex_list = ["excluded" for i in range(len(div_flds))]
div_insert_sql = sql.SQL("""INSERT INTO
                stock_div({})
            VALUES
                ({})
            ON CONFLICT
                (sd_line_id)
            DO UPDATE
                SET
                    ({}) = ({})
            RETURNING
                sd_line_id
            """).format(sql.SQL(", ").join(map(sql.Identifier, div_flds)),
                        sql.SQL(", ").join(map(sql.Placeholder, div_flds)),
                        sql.SQL(", ").join(map(sql.Identifier, div_flds)),
                        sql.SQL(", ").join(map(sql.Identifier, ex_list,
                                               div_flds))
                        )
cur.execute(div_insert_sql, div_data)

Uses the dynamic capabilities of the sql module to build a query. The ex_list is just a list of 'excluded' strings equal in number to the number of the fields that are joined with field names later to create excluded.field_name .