Posts from @graphdatabase on Threads

Snowflake let's you roll your own Aggregate function using the ACCUMULATE Lambda Function. ACCUMULATE first initializes an aggregation context, then iterates through each non-NULL input value to update it. If needed, Merge combines partial contexts from parallel processing, and Terminate uses the completed context to return the final aggregate result. #Snowflake http://quettabyte.net/gist.php?id=IanTI

Snowflake introduces ACCUMULATE, a way to build custom aggregate functions entirely in SQL without external code. It uses four lambda steps—initialize, accumulate, combine, and terminate—to define how the aggregation works. You can create aggregates like PRODUCT or geometric mean directly in #SQL. This eliminates the need for Python, JavaScript, or other languages and makes complex aggregations flexible and scalable. http://quettabyte.net/gist.php?id=jnvj9

WITHOUT OVERLAPS temporal constraint in #Postgres 19 prevents insertion of the overlapping date/time ranges. http://quettabyte.net/gist.php?id=5HkFD

To define Level of Detail metrics in Snowflake Semantic Views you can now use the PARTITION BY EXCLUDING: http://quettabyte.net/gist.php?id=Azd4e