feat: CREATE CATALOG

[BohuTANG]

    > If so, we can do it as a hive table in databend: new catalog and new table engine.

Using a catalog in hive way is not good for me. We should allow users use to configure the catalog for the iceberg. For example, specify the iceberg location by SQL instead of static config.

In this way, we need to introduce a create catalog like:

CREATE CATALOG my_icerberg TYPE='iceberg' CONNECTION=(URL='s3://my_bucket')
SELECT * from my_iceberg.db.table;

Originally posted by @Xuanwo in #8215 (comment)

[Xuanwo]

cc @sandflee for comments.

[BohuTANG]

I am not sure what the statement on the hive table to replace the static config, could you give an example? @dantengsky @Xuanwo

[Xuanwo]

I am not sure what the statement on the hive table to replace the static config, could you give an example? @dantengsky @Xuanwo

The config in databend-query for hive only have two fields:

#[derive(Clone, Debug, PartialEq, Eq, Serialize, Deserialize, Args)]
#[serde(default)]
pub struct HiveCatalogConfig {
    #[clap(long = "hive-meta-store-address", default_value = "127.0.0.1:9083")]
    pub meta_store_address: String,
    #[clap(long = "hive-thrift-protocol", default_value = "binary")]
    pub protocol: String,
}

we can replace it by:

CREATE CATALOG my_hive
  TYPE='hive'
  CONNECTION = (URL='<hive-meta-store>' THRIFT_PROTOCOL=BINARY);
SELECT * FROM my_hive.db.table;

[Xuanwo]

We can defer the hive support after we confirm that it works well on the iceberg.

[Xuanwo]

I'm working on an RFC for this issue.