Create tmp tables with generated unique name.

[binhnq94]

Describe the feature

A clear and concise description of what you want to happen.
With the following configuration:

{{
    config(
        materialized = "incremental",
        file_format = "parquet",
        partition_by = ["created_date"],
        incremental_strategy = "insert_overwrite",
    )
}}

I expected that we could run dbt run -s my_model in parallel with multiple different created_date values passed to vars, for instance, --vars "created_date: 2023-01-01", --vars "created_date: 2023-01-02", ...
However, with the current version 1.7.0, although it seems impossible, dbt-glue will raise an error:

AnalysisException: Table or view 'my_table_tmp' not found in database 'my_db'

For example, if we first run dbt run -s my_model --vars "created_date: 2023-01-01", dbt-glue will create my_model_tmp. Then, if we subsequently run dbt run -s my_model --vars "created_date: 2023-01-02", dbt-glue will attempt to drop my_model_tmp first. This leads to the first run failing as described above.

Describe alternatives you've considered

For each run, need to create tmp table with unique name.

Additional context

Please include any other relevant context here.

Who will this benefit?

Help run incremental model with insert_overwrite strategy in parrallel.

Are you interested in contributing this feature?

I am new to dbt-glue, so I hope for a mentor to guide me through it.