Hi 

I am trying to read data from BQ which is working very slow. I think if I break 
fusion just after reading by calling a ReShuffle , then probably it will 
increase throughput of reads. To execute Reshuffle I need to add a key to BQ 
rows. My table has a billion of rows.

How can I choose a better distribution key , so that a single key gets applied 
to many rows and hopefully after reshuffle it breaks the fusion and starts 
reading data in parallel


Thanks
Aniruddh

Reply via email to