vfrank66

CREATE OR REPLACE FUNCTION damerau_levenshtein(s1 TEXT, s2 TEXT)
        RETURNS INT AS $$
        DECLARE
            s1_len INT := LENGTH(s1);
            s2_len INT := LENGTH(s2);
            d INT[][];
            i INT;
            j INT;
            cost INT;
        BEGIN

vfrank66

CREATE OR REPLACE FUNCTION damerau_levenshtein(source TEXT, target TEXT)
RETURNS INT AS $$
DECLARE
    source_len INT := LENGTH(source);
    target_len INT := LENGTH(target);
    distance INT[][];
    largest_source_chr_matching JSONB := '{}';
    largest_target_chr_matching INT;
    inf INT;
    i INT;

vfrank66

"""
This script migrates databases/tables in AWS Glue backed by S3 data. It assumes you are migrating databases to a new
S3 bucket, it also assumes the data is stored in s3://<bucket>/<database name with standardization across dbs>/<tables>/<optional partitions/data.<ext>

bucket name/
    database1/
        tableA/
        tableB/
    database2/
        table1/

vfrank66

"""AWS DataSync an aws service to move/copy large amounts of data."""
import logging
import os
from typing_extensions import Literal

import boto3
import tenacity
from botocore import waiter
from botocore.exceptions import WaiterError

vfrank66

import logging

from py4j.protocol import Py4JJavaError
from pyspark import SparkContext

logger = logging.getLogger(__name__)


class ServiceApiError(Exception):
    """ServiceApiError exception."""