There are 2 file on my github , you can try it .
test8 with more table than test9, but they are almost same structure.
I have no idea ,why test9 is normal, but test8 is not.
From your Schema, I see the different tables are having same structure.
Is this a test schema provided or an actual schema. Though regardless of the schema details, Liquibase should have worked.
As you mentioned, the test9 works, can you check once by stepwise adding one table at a time in the file. So let’s say on first run, your file has 1st table. On second run, your file have first table and newly added second table, and so on.
Here are my SQL schema, there some Chinese characters , but I wonder know , why my alter always shown is
From this, I see the column names are getting populated that many times as of the number of tables you have in the changeset file. I’m unclear how your test9 works?