To change the PI of a table you will need to recreate the table, But before making changes, make sure to analyse the demographics of the candidate PI columns. Once you have choosen a good PI, then you can recreate the table definition with new PI, and reload it, otherwise you can copy data to some temporary table, recreate the table with new PI and restore the data back to it.
If it is staging table you can create a NOPI table in TD 13.10 or later releases.
You can use hashamp....hashbucket.....hashrow for a field and see how it is distributed per amp and how the skewedness is.
Once you get the best, then you can proceed with recreating the table with the new PI as per your requirement and convenience.
I have few more questions.
1) How to identify the columns to compress and the columns which are not to be compressed?
2) What is the procedure to compress a column?
3) Please provide the script to compress a column and to identify the compression has been done.
Please provide the answer to my queries as earlier as possible
You can analyse the distinct values in columns, less the distinct values, more the column is a candidate for compression.
For example, Gender, country, Codes etc,
To make a column compressible, you need to add the COMPRESS key word in the column definition.
You can find more on Compression here: