The importance of software testing in enterprise system development is growing more and more from cost reduction, quality improvement and time-to-market of their enterprise services. Especially in recent years, companies with the spread of API economy trend has been faced with the need to publish their services via public API-Gateway as Restful-API services. Simultaneously, there is a need for software testing for a part of migrated function from a legacy monolithic architecture to microservice one. In this situation, we have a concern how far range implementation is influenced by getting rid of migrated function's implementation, new function as microservice and isolated data on cloud container. Despite existing several test case extraction methodologies, these are not enough to reveal rage of influence in this case. Therefore, we propose a new test case extraction methodology for a part of system modification and system transition like migrating to microservices from a monolithic system. Our approach focuses on the interaction between function and data to extract test cases from only influenced ranges. On that account, we leverage improved 'Impact Data All Used' method as 'Code-Based - Impact Data All used.' Moreover, we apply graph mining techniques to extracted test cases for reducing the number of test cases efficiently. As a result of this study, by exhaustively searching CRUD operations in source codes level, clarified that it is possible to extract dependencies between functions and data as test cases which are not able to be detected by previous study's IDAU method. Moreover, we suggest the possibility of a test case reduction model by Bonachich Power Centrality and Link-Community analysis.