CustomChange is for implementing a refactoring not currently supported, like User Defined Datatypes.
For this case, I would recommend having the CSV file simple match the table your are importing the data into. It would be easier, I believe, to change the CSV file to match then to creat what you are suggesting.
Hi !
I have a set of testdata (csv) and during load I would like to manipulate some fields (e.g.: dates).
Using CustomChange I must implement all the csv->jdbc statement steps, while LoadDataChange class has already implemented the most of it, and also the parameter handling (path, relative, target table etc)
I wish I could override the CsvReader user by LoadDataChange class to manipulate the read fields (String[]) !
Is it possibble to somehow use my own implementation of LoadDataChange and use the same xml config + custom parameter easily ?
I dont know what CustomChangeWrapper made for, maybe something like this ?
Thanks
Yes indeed, but we want the test writer to permute the data (test)runtime to avoid store hundreds of loader files in the project and repository…
I wrote a working solution as a liquibase extension overriding the LoadDataChange class, extending the <loadData …> tag with an extra attribute !
Maybe it worth to share here when it is totally ready …
