I have a list of users that only administrators can see (= multiple reads). This list also displays the number of users in the data warehouse. Since the list can grow larger than 1000, I first thought to avoid the normal count () and use counter fragments instead.
However, the problem is that administrators also have access to various search filters (in the graphical interface), for example, only for viewing male and female users, etc. It is important that the count reflect these filters so that they can receive the number of female users, male users, and many other combinations.
Because of this, closed counters and high concurrency counters without splinters do not seem like a good idea, because I will need to create a counter for each combination of search filters.
Should I just create a loop of count () methods, like the one described here , or is this a very bad practice? How can I do it differently?
Please note that this counter is for the admin interface and has a very limited number of reads. This is indeed the case when I would like to sacrifice some reading speed for flexibility and accuracy. Although it should be able to grow above 1000, it should not grow above 10 000.
source
share