Andmete migratsiooni testimine
Abstract
Infosüsteemide uuendamise ajajärgul puutuvad infosüsteemide arendajad üha enam kokku vajadusega tõsta (migreerida) vana infosüsteemi andmed uue infosüsteemi andmebaasi. Magistritöös selgitatakse migratsiooni testimist ning leitakse võimalusi testide koostamise ja testimise lihtsustamiseks (automatiseerimiseks). Testimise automatiseerimise vajadusest lähtuvalt käsitletakse migratsiooni testi päringute komplektina, kus igasse komplekti kuulub üks isevalideeruv test (päring) ja vähemalt üks vigaseid kirjeid tagastav abipäring. Sellise lähenemisviisiga saab testide käivitamisel kohese ülevaate migreeritud andmete seisust ja vigaste andmete (kirjete) detailvaate saamiseks ei ole vaja kirjutada SQL lauseid. Seega muutub vigade analüüsimine efektiivsemaks, sest väheneb täiendavate päringute koostamise vajadus. Testide koostamise automatiseerimise eelduseks testide sarnasuse alusel testide tüüpidesse jagamine. Töös defineeriti 16 testi tüüpi ja iga testi tüübi jaoks koostati testi päringu mall (template) ja abipäringute mallid (näidispäringud) ning selgitati testi metaandmete vajadust. Lisas toodud testide tüüpide kirjeldused on näidiseks migratsiooni testijatele. Testide kirjelduste alusel saab arendada ka testide päringute koostamise generaatori.
In the era of information system upgrades, information system developers are increasingly confronted with the need to upgrade (migrate) the old information system data into the database of the new information system. Master's thesis explains migration testing and finds ways to simplify (automate) test design and testing. Based on the need for automation testing, the migration test is considered as a set of queries where every element include one self-validating test (query) and at least one auxiliary query that returns invalid entries. With this approach, one can instantly view the status of migratory data when one runs the tests, and no need to write new SQL statements to get a detailed view of the incorrect data (records). Thus, the analysis of errors becomes more efficient as the need for new additional queries is reduced.The prerequisite for automation of test preparation is the division of tests into test types based on similarity of tests. In this work, 16 test types were defined, and a template for the test query and auxiliary query templates (pattern queries) were prepared for each type of test and the need for test metadata was explained. The descriptions of the types of tests in the Appendix are an example of migration testing. A test generator can also be developed based on test descriptions.
In the era of information system upgrades, information system developers are increasingly confronted with the need to upgrade (migrate) the old information system data into the database of the new information system. Master's thesis explains migration testing and finds ways to simplify (automate) test design and testing. Based on the need for automation testing, the migration test is considered as a set of queries where every element include one self-validating test (query) and at least one auxiliary query that returns invalid entries. With this approach, one can instantly view the status of migratory data when one runs the tests, and no need to write new SQL statements to get a detailed view of the incorrect data (records). Thus, the analysis of errors becomes more efficient as the need for new additional queries is reduced.The prerequisite for automation of test preparation is the division of tests into test types based on similarity of tests. In this work, 16 test types were defined, and a template for the test query and auxiliary query templates (pattern queries) were prepared for each type of test and the need for test metadata was explained. The descriptions of the types of tests in the Appendix are an example of migration testing. A test generator can also be developed based on test descriptions.