Quantcast
Channel: Teradata Forums - Database
Viewing all articles
Browse latest Browse all 14773

Order of partition columns in MLPPI - forum topic by Santanu84

$
0
0

Hi All

 

May be I have posted this question before in this forum, but I am reposting it because that time I did not get any reply. Now I have been asked in a technical meeting to suggest a solution. But I am still not sure how to decide this, as my understanding is still in grey area.

 

"What is best to consider while deciding the order of partition columns in MLPPI?"

 

So far what I have read very frequesntly in PDFs and web links is;

 

a) There are more contiguous partitions to scan and skip at the highest level.

b) You should consider specifying partitioning expression with greatest number of partitions at lowest level.

 

 

I am little confused and not able to figure out the gist from above 2 statements.

 

1. 

If in RANGE partition COL1 is creating 50 partitions and COL2 is creating 10 partitions, 

Shall I use COL1 at the higher level of MLPPI because it is getting more number of partitions than COL2?

Or this will not at all impact the order of partition hierarchy?

 

2.

If a data demographic says, COL1 has 50000 typical rows/value and COL2 has 1000 distinct values then,

Shall I use COL1 at higher level and COL2 at lower level or it should be the opposite?

What is the logic behind this decision?

 

Actually, my understanding is not clear about the order of partition columns in MLPPI.

 

If someone kindly explains this or shares any document or web link where it is depicted, then it will be really helpful. I need to know this answer very urgently.

 

Thanking You

Santanu

 

Forums: 

Viewing all articles
Browse latest Browse all 14773

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>