Hi, I have a question about time-related columns in the default StarTree index configuration. The documentation says:

Here we assume that time columns will be included in most queries as the range filter column and/or the group by column, so for better performance, we always include them as the last elements in the dimensionsSplitOrder

How does putting time columns as the last elements improve performance? Is it related to the number of nodes that need to be processed to solve a given query?

@User If a query contains a filter time column, and the segment doesn't fully match/not match the time filter, in order to use star-tree, the time column needs to be included in the star-tree dimensions. In most queries, time filter is a range, so split on it early might cause too many branches in the star-tree. Based on these concerns, we put time as the last element in the

dimensionsSplitOrder

Makes sense, thank you