Github user jingyimei commented on the issue:
https://github.com/apache/madlib/pull/325
@iyerr3
> I'm getting confused here, so maybe an example would help.
>
> My understanding of @orhankislal 's comment was that say `page_rank` and
`bfs` both use the same table called "vertex" then we will have to drop this
table at the beginning of each of the test files. This may then collide with a
user-created "vertex" table - hence this commit.
> My suggestions to solve the problem:
>
> * Use a function-specific name (`"pagerank_vertex"`) instead of a generic
name with (`"vertex"`) no drop by in the beginning.
This seems doesn't solve the potential conflict with user created table.
Besides, in some IC/DC file we drop tables in between and create table with the
same name again. In this case the second drop/creation will fail if user
already created it.
> * Create a single `"vertex"` table (with no drop by statement) and use
that for all graph functions
This may cause some assertion based on returning number of records fail as
I mentioned in the previous comment, and also needs effort of audition and
combining test data.
> * Schema-qualify the data table with the install-check schema name.
>
(not sure if I totally understand this) install-check schema is based on
module, not algo, for example pagerank and hits and sssp use the same IC
schema, and there will be issue between running two sql files for
droping/creating same table.
> Please note I'm not suggesting pooling **all** install-check data into a
single file (at least not as a solution to this problem).
I understood your point. What I was trying to say was multiple algos in the
same module share the same data creation file.
---
