You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@tez.apache.org by "r7raul1984@163.com" <r7...@163.com> on 2015/06/12 02:23:42 UTC

some advice to tez ui

I think tez ui can add some feature like data governance capabilities such as  - which source table generate the result table,provide lineage search for dag.


r7raul1984@163.com

Re: Re: some advice to tez ui

Posted by "r7raul1984@163.com" <r7...@163.com>.
And what do you mean lineage search for dag ? I mean user can find ins and outs of the sql  by table name.
For example:
INSERT overwrite TABLE TESTTMP.cust_sku_vis_detl1 SELECT 
      ......
FROM 
DW.fct_ordr_detl detl 
JOIN DW.fct_ordr ordr ON detl.ordr_id = ordr.ordr_id 
AND detl.ds = ordr.ds 
JOIN DW.dim_prod prod ON detl.prod_id = prod.prod_id 
AND prod.cur_flag = 1 
JOIN DW.hier_categ categ ON prod.categ_lvl_id = categ.categ_lvl_id 
AND categ.cur_flag = 1 
WHERE 
detl.ds = '2015-06-01' 
AND ordr.ds = '2015-06-01' 
GROUP BY 
ordr.ds, 
ordr.end_user_id, 
categ.categ_lvl1_id, 
categ.categ_lvl1_name

Tez UI can pick out some sql (dag graph)   when user search for TESTTMP.cust_sku_vis_detl1.




r7raul1984@163.com
 
From: Jeff Zhang
Date: 2015-06-12 08:38
To: user
Subject: Re: some advice to tez ui
This might need hive add some app specific info on the dag/vertex. Actually Tez don't know the concept of table, tez use InputDescriptor/OutputDescriptor to represent the Input/Output which is not restricted to hive table, it could be any data source.
 
And what do you mean lineage search for dag ?

On Fri, Jun 12, 2015 at 8:23 AM, r7raul1984@163.com <r7...@163.com> wrote:

I think tez ui can add some feature like data governance capabilities such as  - which source table generate the result table,provide lineage search for dag.


r7raul1984@163.com



-- 
Best Regards

Jeff Zhang

Re: some advice to tez ui

Posted by Jeff Zhang <zj...@gmail.com>.
This might need hive add some app specific info on the dag/vertex. Actually
Tez don't know the concept of table, tez use
InputDescriptor/OutputDescriptor to represent the Input/Output which is not
restricted to hive table, it could be any data source.

And what do you mean lineage search for dag ?

On Fri, Jun 12, 2015 at 8:23 AM, r7raul1984@163.com <r7...@163.com>
wrote:

>
> I think tez ui can add some feature like data governance capabilities such
> as  - which source table generate the result table,provide lineage search
> for dag.
> ------------------------------
> r7raul1984@163.com
>



-- 
Best Regards

Jeff Zhang

Re: some advice to tez ui

Posted by Gopal Vijayaraghavan <go...@apache.org>.
> I think tez ui can add some feature like data governance capabilities
>such as  - which source table generate the result table,provide lineage
>search for dag.

Tez is an execution engine and the wrong layer to collect this sort of
information.

If those are your actual requirements (for hive), checkout Apache Atlas -
http://wiki.apache.org/incubator/AtlasProposal

Cheers,
Gopal