Kuan-Po Tseng

Currently we don't have columns in view. this still fail since accio.test not removed ```SQL SELECT accio.test.col_1 FROM accio_view; ```

ERROR: com.google.cloud.bigquery.BigQueryException: Column name suppkey is ambiguous at [54:6] sql ``` SELECT * FROM SupplierRevenue ``` json ``` { "catalog": "canner-cml", "schema": "tpch_tiny", "models": [ { "name": "Orders", "refSql": "select...

Describe only need table schema, while we execute whole sql, this. brings redundant performance overhead To solve this, we could: - add limit 1 for each sql - use describe...

Add unit tests to validate sql for `createPgClass` `createPgType` `createPgAmTable` .etc.

duckdb and postgresql both use array[1:3] to slice array.

https://github.com/Canner/canner-metric-layer/issues/284#issuecomment-1590787241

https://github.com/Canner/canner-metric-layer/pull/180#discussion_r1128976808

StatementAnalyzer#visitTable should analyze if table is in WITH-Query, if so, it doesn't need to use `toCatalogSchemaTableName`

### Summary this is part of https://github.com/harvester/harvester/issues/5288, after https://github.com/harvester/harvester/pull/5386 merged to master, error message updated upon each retry failure. Before this pr, error message pops up only when upgrad CR...