For Kusto queries on very large tables, which of the following saves more time: (A) projecting to only keep columns of relevant data first, or using where to filter out rows first?
For Kusto queries on very large tables, which of the following saves more time: (A) projecting to only keep columns of relevant data first, or using where to filter out rows first?
You want to do both and the order does not matter since the query optimizer will arrange it in the most optimal way.
As a side note, if your query summarizes the data, then there is no need for an explicit project, only relevant columns will be used.