Github user hyunsik commented on the pull request:
https://github.com/apache/tajo/pull/273#issuecomment-65762633
It's really great job. I briefly investigated it in order to give good
reviews. I'm still reviewing it. I'll give more comments soon.
Now, I have one suggestion or discussion agenda related to this issue. It
would be nice if ultimately we have a direction like **INFORMATION SCHEMA**
(http://en.wikipedia.org/wiki/Information_schema). You may already know this
concept more than me. It is required for Tajo to follow ANSI SQL standard.
BTW, I don't think that we have to do it in this issue. We can commit your
current work now, and mark it as an experimental and evolving feature, Then, we
change it later if other committers agree with my suggestion. I'd also like to
help this work.
The following is my suggestion for the future. There may be two approaches
to achieve ```INFORMATION SCHEMA```.
One is that we change current schemas to follow ```INFORMATION SCHEMA```.
But, this approach requires lots of burden. Also, it may be hard to optimize
table schemas for our system characteristics.
Another approach is to provide virtual tables for ```INFORMATION SCHEMA```.
Actually, we don't need to have catalog schemas which are physically equivalent
to ```INFORMATION SCHEMA```. For example, Postgresql uses different schemas to
maintain its catalog, but it provides a virtual table to show information
schema. I think that we can do like them.
* information schema of pgsql -
http://www.postgresql.org/docs/current/interactive/information-schema.html
* internal catalog schema of pgsql -
http://www.postgresql.org/docs/current/interactive/catalogs.html
How do you think about my future direction?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
