Hi Ben, 
There are several ways of working with arrays in Drill.  The repeated_count( 
<array>) function returns the number of items in an array, and the 
repeated_contains(<array>, ’needle') returns true/false if an array contains a 
given item.  You might want to take a look at this page from the docs as well: 
https://drill.apache.org/docs/lesson-3-run-queries-on-complex-data-types/

https://drill.apache.org/docs/repeated-count/
https://drill.apache.org/docs/repeated-contains/

Best,
— Charles



> On Jan 9, 2019, at 05:37, [email protected] 
> <[email protected]> wrote:
> 
> I think I write too quickly, since the Parquet format seems to allow the use 
> of "list" type. 
> So I do not understand why there is no easy manipulation of array in DRILL.
> 
> ----- Mail original -----
> Hi,
> 
> Although it's possible to work with ARRAY in DRILL :
> SELECT columns[0], columns[1]... FROM ...;
> 
> I don't find any array function to know length or dimension of array or any 
> other useful fct
> 
> I don't find any possibility to "build" my array with different data like :
> SELECT ARRAY[data1, data2, data3] FROM ...;
> 
> Aggregate function that build array doesn't exists
> Example : SELECT mycolumn1, array_agg(mycolumn2) FROM mytable GROUP BY 
> mycolomn1;
> 
> Or would I miss something ?
> 
> I know that array type doesn't exists for Parquet files but these capacities 
> will be very useful
> and the limitation of Parquet should not limit the possibilities of DRILL.
> 
> Appreciate any info or return on this subject ?
> 
> Regards,

Reply via email to