You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Many programs need to know the number of results rather than the individual results. Currently, VLog4j can only get this number by iterating through all query results to count them.
Instead, Vlog should have a direct method to return the number of results of a query. Like the query answering method, it should support counting with and without nulls. In many cases, this number should be much more efficiently retrievable by VLog. summing up the total size of IDB tables for a predicate should be easy. Indeed, it would be good to have a special method to count all facts for a predicate (no full-fledged query). Counting all facts that contain no nulls at all could also be easy if each IDB table would store this number (if it can be obtained cheaply when building the table).
For more specific queries (with constants or repeated variables), iteration would still be needed, but it is presumably faster if done in C++ directly rather than by sending all results as strings to Java.
The text was updated successfully, but these errors were encountered:
Many programs need to know the number of results rather than the individual results. Currently, VLog4j can only get this number by iterating through all query results to count them.
Instead, Vlog should have a direct method to return the number of results of a query. Like the query answering method, it should support counting with and without nulls. In many cases, this number should be much more efficiently retrievable by VLog. summing up the total size of IDB tables for a predicate should be easy. Indeed, it would be good to have a special method to count all facts for a predicate (no full-fledged query). Counting all facts that contain no nulls at all could also be easy if each IDB table would store this number (if it can be obtained cheaply when building the table).
For more specific queries (with constants or repeated variables), iteration would still be needed, but it is presumably faster if done in C++ directly rather than by sending all results as strings to Java.
The text was updated successfully, but these errors were encountered: