The current ReplacementDataSet is too limited. It's too hardcoded around strings and maps of replacements.
The concepts of wrapping dataset/tables to replace values in them, and actually replacing values, should be separated.
More interesting replacements are possible too:
- Replace value with current time
- Replace value prefixed "blob:" with BLOB content read from file/URL
- Replace value prefixed "clob:" with text read from file/URL
- Replace value prefixed "script:" with result of scripting function
... etc.
dbunit: 5c487f4588dde20ccd40cbec
dbunit: 5c51c8c4860ca8723ac53c27