org.apache.spark.sql.catalyst.expressions.aggregate.ApproximatePercentile
PercentileDigest
Related Doc:
package ApproximatePercentile
org.apache.spark.sql.catalyst.expressions.aggregate.ApproximatePercentile
underlying probabilistic data structure QuantileSummaries.
An internal flag from class QuantileSummaries to indicate whether the underlying quantileSummaries is compressed.
Insert an observation value into the PercentileDigest data structure.
Returns the approximate percentiles of all observation values at the given percentages.
Returns the approximate percentiles of all observation values at the given percentages. A percentile is a watermark value below which a given percentage of observation values fall. For example, the following code returns the 25th, median, and 75th percentiles of all observation values:
val Array(p25, median, p75) = percentileDigest.getPercentiles(Array(0.25, 0.5, 0.75))
In-place merges in another PercentileDigest.
Returns compressed object of QuantileSummaries
PercentileDigest is a probabilistic data structure used for approximating percentiles with limited memory. PercentileDigest is backed by QuantileSummaries.