Please look into logs for more details.
Not sure why you see these errors but Drill can perfectly query singe files, 
subset of files and directories.

select * from dfs.tmp.`*.parquet` limit 4;
select * from dfs.tmp.`0_0_0.parquet`;

Kind regards,
Arina

> On Feb 4, 2020, at 7:10 PM, Nitin Pawar <[email protected]> wrote:
> 
> as the error says .. it expects a directory to query
> also the document has not been modified for more than 3 years so not sure
> if it up to date
> 
> On Tue, Feb 4, 2020 at 10:30 PM Vishal Jadhav (BLOOMBERG/ 731 LEX) <
> [email protected]> wrote:
> 
>> I was following the help pages from here.
>> https://drill.apache.org/docs/querying-parquet-files/
>> As per it, I can query an individual parquet file, why is it failing with
>> the 'not a directory' error.
>> 
>> 
>> From: [email protected] At: 02/04/20 11:28:25To:  Vishal Jadhav
>> (BLOOMBERG/ 731 LEX ) ,  [email protected]
>> Subject: Re: Drill + parquet
>> 
>> Parquet is default file format for apache drill
>> so you do not need to give a parquet file for a drill query. Instead give
>> the folder path which contains the files.
>> 
>> eg: select * from hdfs_storage>.<workspace>.`folder1` will query all the
>> parquet files in folder1
>> 
>> On Tue, Feb 4, 2020 at 9:55 PM Vishal Jadhav (BLOOMBERG/ 731 LEX) <
>> [email protected]> wrote:
>> 
>>> Hello Drillers,
>>> 
>>> Need some help with the hdfs + parquet files.
>>> 
>>> I have configured the HDFS storage with parquet & csv format plugins.
>>> 
>>> I can query the - <hdfs_storage>.<csv_ws_name>.`*.csv` correctly. Also, I
>>> have a similar directory structure for the parquet files (in a different
>>> directory), But, not able to query it.
>>> 
>>> Show files works fine.
>>> (1) The following query works fine -
>>> show files from <hdfs_storage>.<workspace>
>>> 
>>> (2) select * from <hdfs_storage>.<workspace>.`*.parquet` limit 4
>>> Fails with -
>>> org.apache.drill.common.exceptions.UserRemoteException: SYSTEM ERROR:
>>> NoSuchElementException
>>> 
>>> (3) select * from hdfs_storage>.<workspace>.`xyz.parquet`;
>>> fails with -
>>> org.apache.drill.common.exceptions.UserRemoteException: SYSTEM ERROR:
>>> RemoteException:/path/xyz.parquet (is not a directory)
>>> 
>>> Please let me know, if I am doing something wrong here.
>>> 
>>> Thank you!
>>> - Vishal
>> 
>> 
>> --
>> Nitin Pawar
>> 
>> 
>> 
> 
> -- 
> Nitin Pawar

Reply via email to