apache/datasketches-cpp

Use non-colliding family id in count-min

Closed this issue · 0 comments

In order to have proper binary compatibility between C++ and Java and allow valid deserialization checks, we need to ensure we have a common family_id namespace. In Java it exists in a single enum class: https://github.com/apache/datasketches-java/blob/master/src/main/java/org/apache/datasketches/common/Family.java

The count-min sketch's family_id currently collides with Java's Alpha variant of the theta sketch: https://github.com/apache/datasketches-cpp/blob/master/count/include/count_min.hpp#L322
Let's assign count-min an id of 18, and then make a PR against the Java repo to reserve that value in the Family enum.