[
https://issues.apache.org/jira/browse/DRILL-5556?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16028682#comment-16028682
]
Aman Sinha commented on DRILL-5556:
-----------------------------------
I am not sure if it is an invalid SQL. Postgres certainly allows it:
{noformat}
tpcds=# select * from tt1 ;
a1 | b1
----+----
10 | 20
(1 row)
tpcds=# select *, * from tt1 ;
a1 | b1 | a1 | b1
----+----+----+----
10 | 20 | 10 | 20
(1 row)
{noformat}
Only when you reference the columns in some way, e.g from an outer query block
then it will given an error:
{noformat}
tpcds=# select a1 from (select *, * from tt1) as d;
ERROR: column reference "a1" is ambiguous
LINE 1: select a1 from (select *, * from tt1) as d;
{noformat}
> Drill allows "SELECT *, *" query
> --------------------------------
>
> Key: DRILL-5556
> URL: https://issues.apache.org/jira/browse/DRILL-5556
> Project: Apache Drill
> Issue Type: Bug
> Affects Versions: 1.10.0
> Reporter: Paul Rogers
> Priority: Minor
>
> Configure the CSV storage plugin to not read the header row. Create a simple
> CSV file:
> {code}
> 10,foo,bar
> {code}
> Issue the following nonsensical query:
> {code}
> SELECT *, * FROM `dfs.data`.`example.csv`
> {code}
> Get the following results:
> {code}
> columns,columns0
> ["10","foo","bar"],["10","foo","bar"]
> {code}
> Expected an error as using two unqualified *'s is not valid SQL.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
