You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@arrow.apache.org by Jin Hai <mi...@sina.com> on 2017/12/24 08:48:04 UTC

Is there any way to get data of numeric array like maximum/minimum value?

Hi,
	I am going to check the arrow array data range, if the range doesn’t include my data, I will scan it any more. Do we have such way to get this kind of data?

Thanks

Best Regards
Jin Hai

Re: Is there any way to get data of numeric array like maximum/minimum value?

Posted by Wes McKinney <we...@gmail.com>.
We are still in the early stages of defining such APIs. See

https://github.com/apache/arrow/tree/master/cpp/src/arrow/compute

where this code is being written. We would be glad of any additional
work or input on the design for the computation APIs

- Wes

On Tue, Jan 2, 2018 at 10:43 AM, Jin Hai <mi...@sina.com> wrote:
>
> Cool, do you have any idea how to implement the function in Arrow? Such as create a JIRA ticket and describe the idea?
>
>> On 2 Jan 2018, at 11:37 PM, Wes McKinney <we...@gmail.com> wrote:
>>
>>
>> Hi Jin,
>>
>> If you are reading Parquet files, you can look at the column
>> statistics for a row group using the C++ API. We definitely plan to
>> implement functions to compute reduction-based statistics like minimum
>> and maximum from data already in an Arrow array in C++, but it has not
>> been implemented yet.
>>
>> - Wes
>>
>> On Sun, Dec 24, 2017 at 3:48 AM, Jin Hai <mi...@sina.com> wrote:
>>>
>>> Hi,
>>>        I am going to check the arrow array data range, if the range doesn’t include my data, I will scan it any more. Do we have such way to get this kind of data?
>>>
>>> Thanks
>>>
>>> Best Regards
>>> Jin Hai
>>>
>>>
>>
>>
>
>

Re: Is there any way to get data of numeric array like maximum/minimum value?

Posted by Jin Hai <mi...@sina.com>.
Cool, do you have any idea how to implement the function in Arrow? Such as create a JIRA ticket and describe the idea?

> On 2 Jan 2018, at 11:37 PM, Wes McKinney <we...@gmail.com> wrote:
> 
> 
> Hi Jin,
> 
> If you are reading Parquet files, you can look at the column
> statistics for a row group using the C++ API. We definitely plan to
> implement functions to compute reduction-based statistics like minimum
> and maximum from data already in an Arrow array in C++, but it has not
> been implemented yet.
> 
> - Wes
> 
> On Sun, Dec 24, 2017 at 3:48 AM, Jin Hai <mi...@sina.com> wrote:
>> 
>> Hi,
>>        I am going to check the arrow array data range, if the range doesn’t include my data, I will scan it any more. Do we have such way to get this kind of data?
>> 
>> Thanks
>> 
>> Best Regards
>> Jin Hai
>> 
>> 
> 
>

Re: Is there any way to get data of numeric array like maximum/minimum value?

Posted by Wes McKinney <we...@gmail.com>.
Hi Jin,

If you are reading Parquet files, you can look at the column
statistics for a row group using the C++ API. We definitely plan to
implement functions to compute reduction-based statistics like minimum
and maximum from data already in an Arrow array in C++, but it has not
been implemented yet.

- Wes

On Sun, Dec 24, 2017 at 3:48 AM, Jin Hai <mi...@sina.com> wrote:
>
> Hi,
>         I am going to check the arrow array data range, if the range doesn’t include my data, I will scan it any more. Do we have such way to get this kind of data?
>
> Thanks
>
> Best Regards
> Jin Hai
>
>