All Forums Database
kamlesh114 7 posts Joined 05/15
26 Jul 2016
What is difference between Redistribution of rows and Duplication of small table in AMPs

Related to join strategy used by TD optimzer.
When we join table and see explain plain then  it show that rows are redistributed/duplicated on all AMPs based on join startegy used by Optimizer to get the final result.
Please correct me..
1.Redistribution - Rows from one or both table are redistributed to all the AMPs based on hash code.
2.Duplicate table - it duplicate the entire table on all the MAPs to make join . (normally optimizer choses to duplicate the small table on all AMPs).

26 Jul 2016

Yes, that is correct at a high level.
Redistribution will be based on hash code of the joining column(s)

kamlesh114 7 posts Joined 05/15
26 Jul 2016

Thank you Sachin.
2nd point is also correct ? where TD duplicate the entire table (small table )on all AMPs.
And also please tell me when the rows are redistributed or duplicated then where does TD actually moves the rows ..
If in spool space then which spool space..
AMPs spool space or there is other spool space ?  

M.Saeed Khurram 544 posts Joined 09/12
27 Jul 2016

Teradata has only one spool space, the rows are duplicated / redistributed to spool space.


You must sign in to leave a comment.