Comparison
Hive-style partitioningvsPartition pruning
Hive-style partitioning
the partition value is encoded in the folder name as dt=2026-01-14, and the column does not exist inside the files at all.
Encoding partition values in the object key path, so a reader infers them from the directory structure. It is simple, portable and universally understood, and it is why so much of the lake looks like a folder tree. Its weaknesses are what the table formats fix: renaming a partition scheme means rewriting every path, listing is expensive at scale, and the partition value is a string that must be parsed rather than a typed column.
Full entry →Partition pruning
you filtered on the partition column and the engine opened four folders instead of nine hundred.
Skipping whole partitions because the query's filter proves they cannot contain matching rows. It is the coarsest and by far the most valuable form of skipping, which is why partition column choice is the most consequential physical decision in a warehouse. It fails silently when the filter is on a derived expression the engine cannot see through, so a function wrapped around the partition column can turn a two-second query into a full scan.
Full entry →