I have been looking at increasing the performance of some tables that have been uploaded to a redshift database. One of the columns in a table has many delimited values which need to be split out and queried at runtime. The table has around 30 million rows and there would be on average 10 delimited values in this particular column.
Coming from an OLTP background my instinct is to normalize these values out to another table. However, I have seen various posts that suggest that this is not the way to go. Looking for some advice