An analyst runs transaction across a week of proxy logs for every city employee to find rare destinations. Why is that an efficiency anti-pattern, and what is the better pattern?
Select an answer to reveal the explanation.
Short Explanation
transaction across a week of city proxy logs is like stacking every commuter's whole trip into RAM. If the hunt only needs grouping, stats or streamstats by user and time buckets does the job without the memory bill.
Full Explanation
transaction is expensive because it holds related events in memory to build sessions. Running it across a week of proxy logs for every city employee is an anti-pattern when the analytic need is grouping or rare-destination counts. stats or streamstats by user with time buckets is the cheaper replacement. makeresults does not search proxy data, and a leading URL wildcard makes the search worse, not better.