drill-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Kunal Khatua <kkha...@maprtech.com>
Subject Re: Speed up Drill
Date Sat, 03 Sep 2016 00:29:21 GMT
If you are using the JDBC plugin to connect to SqlServer, which is a community storage plugin,
performance might be sub-optimal. 

If you are trying to do a table scan, you also have the possibility that the data being streamed
is very large. Your terminal might be the bottleneck for having to display each row of the
resultset. One suggestion is to use a wrapper SQL of count(*) around this query to see if
that is a possible cause. You should be still scanning all the data, but you wouldn't flood
your SQL terminal.

e,g,

select count(*) from ( <your SQL query> ) ;


On Thu 1-Sep-2016 4:24:27 AM, Khurram Faraaz <kfaraaz@maprtech.com> wrote:
You are doing a scan over a file, and there is no predicate/filter in your
query, which means Drill will have to Scan the entire file.

Also Apache Drill's latest released version is 1.8.0, you may want to try
your query on Apache Drill 1.8.0.

On Thu, Sep 1, 2016 at 11:34 AM, wrote:

> Hello I am using Drill in Window 10 in embedded mode having latest
> version (1.7).I am writing one select Query for Sql Server Database.
> Query is:-
> SELECT CAST(T1.`ID` AS BIGINT) AS `ID`, CAST(T1.`AttributeID` AS BIGINT)
> AS `AttributeID`, T1.`String_Value` AS `String_Value`,
> CAST(T1.`Numeric_Value` AS BIGINT) AS `Numeric_Value`, T1.`MetaData` AS
> `MetaData`, T1.`DataType` AS `DataType`, T1.`CreatedTime` AS `CreatedTime`,
> CAST(T1.`CreatedBy` AS BIGINT) AS `CreatedBy` FROM demoplugin.`dbo`.`ProductValues`
> AS T1
> Result:- It returns 1,113,843 rows in 107.126 Second. ( Too much time)
> My question is is there any way to speed up my drill, so that query
> execute fast.?
>
> Sent from Yahoo Mail. Get the app

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message