Github user orhankislal commented on the issue:
https://github.com/apache/madlib/pull/325
> Also note: madpack is supposed to drop the file-specific schema after
executing each file (see function _execute_per_module_install_dev_check_algo).
Hence, common table names in independent tests within same module are not
supposed to conflict with each other (if you've seen this happen then it
requires investigation).
Oh, I see. The naming convention is somewhat strange. Under the modules
folder, we have a bunch of folders (graph, etc.) but they are not actually
modules, the individual sql files are. The madpack code uses the variable
`module` for the folder name which further muddles the naming. This means
`madlib_installcheck_graph` schema will be created and dropped for each module
in graph. We might want to change it to reflect the actual module name.
I think the reused name issue might be more widespread than we think. I am
pretty sure `abalone` and `houses` datasets are used in multiple modules.
I think removing the `DROP TABLE` statements might work as @iyerr3
suggested. I'll keep the PR open for now to keep the conversation open and
start working on a different branch.
---
