Key in dictionary does not always exist

[jkhairal]

Describe the bug

In the following line of the code:
https:/awslabs/aws-data-wrangler/blob/main/awswrangler/catalog/_create.py#L53 (And/Or Line 96)
If the Glue table actually exists and does not have any tags, it will not return table_inputs["Parameters"] and there will be a Key Error at the above lines.

Environment
Python:

 File "test.py", line 345, in main
    upload_data_to_s3(entries, bucket, database, "ticket_detailed_entries", 'modified_date', log)
  File "test.py", line 201, in upload_data_to_s3
    wr.s3.to_parquet(
  File "/apollo/env/HWEngDSADataIngester/lib/python3.8/site-packages/awswrangler/_config.py", line 417, in wrapper
    return function(**args)
  File "/apollo/env/HWEngDSADataIngester/lib/python3.8/site-packages/awswrangler/s3/_write_parquet.py", line 582, in to_parquet
    catalog._create_parquet_table(  # pylint: disable=protected-access
  File "/apollo/env/HWEngDSADataIngester/lib/python3.8/site-packages/awswrangler/catalog/_create.py", line 262, in _create_parquet_table
    _create_table(
  File "/apollo/env/HWEngDSADataIngester/lib/python3.8/site-packages/awswrangler/catalog/_create.py", line 96, in _create_table
    table_input["Parameters"]["projection.enabled"] = "false"

To Reproduce

Create a simple glue table with no tags, call the following function:

wr.s3.to_parquet(
    df=dataframe,
    path=f"s3://{output_bucket}/{output_s3_prefix}",
    index=False,
    dataset=True,
    mode='overwrite_partitions',
    partition_cols=[partition],
    concurrent_partitioning=True,
    database=database,  # Athena/Glue database
    table=table,  # Athena/Glue table
    dtype=all_types
)

[jaidisido]

Should be fixed by #999 and will be available in next release. Thank you for raising