The tool "Convert Labels to Annotation" creates in the first step all possible annotations. In the second step the labels which are defined under the properties of each feature-class are written into the existed empty annotation. After that we have 50 and more annotations in the table of content, but only 3 annotations are filled. Now it is difficult to get an overview about the annotations in the map. The function "remove empty layers" will not remove empty annotations from the table of content. We want to create masks for the annotations with our own python-script. The problem is, each annotation will be taken into account by our python-script. That is why the process runs very long. In the moment we have to use an additional software / tool to remove empty annotations. The other point is, that the database is very big through the big number of annotations. Our proposal is, that the tool "Convert Labels to Annotation" should first check whether labels have to be written to annatation and if not, then the next feature-class have to check. No empty annotation should be created.