Skip to main content
Version: Candidate-3.4

Aggregate

πŸ“„οΈ percentile_disc_lc

Returns a percentile value based on a discrete distribution of the input column expr. Same behavior as percentiledisc. However, the implementation algorithm is different. percentiledisc needs to obtain all input data, and the memory consumed by merge sorting to obtain percentile values ​​is the memory of all input data. On the other hand, percentiledisclc builds a hash table of key->count, so when the input cardinality is low, there is no obvious memory increase even if the input data size is large.