The following entities must be deployed in order for CLOE to work correctly. In general technical entities like the FileCatalog, file formats, external/internal stages, stored procedure either internal or as output of Exec SQL jobs are assumed to reside in the cloe_dwh schema. The cloe_dwh schema needs to be accsessible via the CLOE technical users default database.
The cloe_dwh schema or why CLOE fails to find stored procedures
CLOE does not use absolute reference notation when executing stored procedures either as the output of an Exec SQL job or as an internal CLOE procedure. Instead it simply executes:
If the cloe_dwh schema is not placed in the database that is defined for the CLOE technical user as default or defined in the connection string, CLOE will not find it. Same is true for the FileCatalog.
All stage tables, ods and/or core tables must also exist once the ETL process starts.
CLOE stores metadata about files already processed or loaded in special table called FileCatalog. This needs to be deployed in the cloe_dwh schema and accessible by the CLOE ETL users.
The following functions must be created and accessible by the CLOE ETL user. If Snowflake is used as a DWH the external/internal stage in the cloe_dwh schema must be accessible by the CLOE ETL user.
CLOE needs this function to retrieve files directly from storage. This is the case when CLOE is not doing the extraction and only processing files provided to it by external tools.
The following paragraphs are mostly directed towards experienced users in Snowflake. If you are unsure how to create stages,
work with file formats and how the snowflake privilege concept works, feel free to reach out to the Cloud-DWH Team.
CLOE needs read/write access to all tables that are part of the ETL process. Besides it needs usage privileges for procedures and lastly access to all stages it is going to retrieve data from. This guide is preliminary and will be updated to be properly aligned with the role concept.