Abstract: A user interface may include one or more visualizations based on queries to a database using query model metadata. Updated metadata of columns of the database may be obtained which may be different from the query model metadata. A set of used columns and a set of unused columns may be determined based on the query model. One or more metadata mismatches may be determined based on a comparison of the query model metadata for the set of used columns and the updated metadata for the set of used columns. The one or more metadata mismatches may be presented in a user interface configured to enable a user to rectify each. One or more selections from the user may be obtained to rectify at least one of the one or more metadata mismatches. The corresponding columns may be deleted or replaced based on the one or more selections.

Abstract: An application embedded within a cloud application may load a data object which accesses information from a database to generate visualizations of data. When the data object is transported from a development system to a production system, the database may not have table entries and views as expected and loading the data object might fail. In such cases, the datasource may be created from metadata. A first metadata call is sent to the database using the application metadata and an indication that a datasource was not found is received from the database. Query metadata is extracted from the application metadata of the data object and a datasource creation call is sent to the database using the extracted query metadata, thereby initiating creation of a datasource artifact in the database. Then, query results are obtained based on the datasource artifact and query results are provided in the cloud application.