Let's say my application is working happily for a year. I wonder if the partitions of a particular table is well balanced. Basically I want to plot the histogram of the distribution of the rowcounts of ALL the partitions in the table.
Granted that this is an anti-pattern. But for the sake to technicallity, I would like to know how this could be done, and of course, let's assume I accept to pay the penalty for that one time table-scan query.
In the relational world, I would do:
SELECT myPartitionKey, COUNT(*) AS rowcount FROM myTable GROUP BY myPartitionKey ORDER BY COUNT(*) DESC;
How can we reproduce the equivalent with Cassandra? Any solution would do, CQL or writing a program.