[
https://issues.apache.org/jira/browse/HAWQ-393?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15287777#comment-15287777
]
Michael Andre Pearce (IG) edited comment on HAWQ-393 at 5/17/16 10:35 PM:
--------------------------------------------------------------------------
So in tableau you can create manual sql instead of using the ui self discovery
feature to connect to a data source.
re Tableau, i suspect they use standard commands via the odbc driver. As i
would expect any odbc compliant product such as excel, powerbi etc.
I typically would expect products like tableau to talk to odbc and its
responsibility of the db to provide odbc compatible drivers that either do this
mapping in their supplied driver or in the server. In your case i believe
currently it is postgres odbc drivers and the data direct odbc that are meant
to be compatible.
Just a bit of background why I'm so interested in this area:
I work for a company that uses tableau heavily connecting to currently more
traditional db's- postgres, oracle, and also a small analytics db veritca - and
we are currently evaluating a better/complimentry sql on hadoop solution where
performance and sql completeness are key issues we have with Hive, our cluster
is a hdp cluster and we adopt open source first architecture policy where
suitable products can be found which for obvious reasons is why hawq is being
looked at, though obviously comparing with other commercial only products also
was (Author: michael.andre.pearce):
So in tableau you can create manual sql instead of using the ui self discovery
feature to connect to a data source.
re Tableau, i suspect they use standard commands via the odbc driver. As i
would expect any odbc compliant product such as excel, powerbi etc.
I typically would expect products like tableau to talk to odbc and its
responsibility of the db to provide odbc compatible drivers that either do this
mapping in their supplied driver or in the server. In your case i believe
currently it is postgres odbc drivers and the data direct odbc that are meant
to be compatible.
Just a bit of background why I'm so interested in this area:
I work for a company that uses tableau heavily connecting to currently more
traditional db's- postgres, oracle, and also a small analytics db veritca - and
we are currently evaluating a better/replacement sql on hadoop solution where
performance and sql completeness are key issues we have with Hive, our cluster
is a hdp cluster and we adopt open source first architecture policy where
suitable products can be found which for obvious reasons is why hawq is being
looked at, though obviously comparing with other commercial only products also
> HAWQ access of hcatalog metadata using \d
> -----------------------------------------
>
> Key: HAWQ-393
> URL: https://issues.apache.org/jira/browse/HAWQ-393
> Project: Apache HAWQ
> Issue Type: New Feature
> Components: Hcatalog, PXF
> Reporter: Goden Yao
> Assignee: Oleksandr Diachenko
> Fix For: 2.0.0
>
>
> As a Hawq user, after enabling hcatalog integration feature,
> I want to be able to display (\d) tables in hcatalog so that I know what
> tables are available and the contents of their schemas.
> *Acceptance Criteria*
> 1. Ability to describe a single table.
> {code}
> \d hcatalog.databaseName.t1
> {code},
> {code}
> \d+ hcatalog.databaseName.t1
> {code}
> 2. Ability to describe the every table in a database of hcatalog:
> {code}
> \d hcatalog.databaseName.*
> {code}
> 3. Ability to list tables that match patterns (only wildcard is supported)
> e.g.
> {code}
> \d hcatalog.databaseName.prefix_*
> {code}
> 4. Ability to list databases in hcatalog:
> {code}
> \d hcatalog.*
> {code}
> 5. Tab autocomplete of items from hcatalog. (optional)
> [1] based on current implementation, we cache the meta data from hcatalog
> within the same transaction.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
