CODEWITHSUNDEEP
google-bigquery
foxwendy
foxwendy

Reputation: 2929

Bigquery performance issue today?

It takes like forever to run a query or a query job with large response today. It just keep running for dozens of minutes without any errors . Those are all queries which I used in the last few days and they all went quite well. One of those is :

select  sub.flag_time as flag_time 
    ,sub.hardware_id as hardware_id   
    ,t1.date_time as date_time   
    ,t1.latitude as latitude   
    ,t1.longitude as longitude   
    ,t1.zip_code as zip_code   
    ,t1.ignition as ignition   
    ,t1.speed as speed  
from 
( 
select flag_time, hardware_id, max(signature) as signature 
from fiona_test.online_actual_tmp  
group each by flag_time, hardware_id 
)as sub 
LEFT JOIN EACH 
(    
select  
     hardware_id 
    ,date_time   
    ,latitude   
    ,longitude  
    ,zip_code   
    ,ignition  
    ,speed 
    ,STRING(file_date_time)+STRING(date_time)+STRING(latitude)+STRING(longitude)+STRING(gps_valid)+STRING(ignition)+STRING(speed)+STRING(gps_reason)+STRING(zip_code) as signature  
FROM TABLE_DATE_RANGE(Geotab_test.gps32_,DATE_ADD('2014-10-01 00:00:00.339',-100,'SECOND'),TIMESTAMP('2014-10-01 23:58:20.339'))   
  GROUP EACH BY hardware_id,date_time,latitude,longitude,zip_code,ignition,speed,signature 
    HAVING (date_time BETWEEN DATE_ADD('2014-10-01 00:00:00.339',-100,'SECOND')+1 AND TIMESTAMP('2014-10-01 23:58:20.339'))  
)as t1 
ON t1.hardware_id = sub.hardware_id and t1.signature = sub.signature

Upvotes: 2

Views: 125

Answers (1)

Jordan Tigani
Jordan Tigani

Reputation: 26617

There is an ongoing shuffler issue affecting queries that use JOIN EACH and GROUP EACH BY. We're currently investigating, it should now be fixed for new queries.

Upvotes: 2

Related Questions