mirror of
https://github.com/elastic/elasticsearch.git
synced 2025-06-29 01:44:36 -04:00
17 lines
1.1 KiB
Markdown
17 lines
1.1 KiB
Markdown
* configurable precision, which decides on how to trade memory for accuracy,
|
|
* excellent accuracy on low-cardinality sets,
|
|
* fixed memory usage: no matter if there are tens or billions of unique values, memory usage only depends on the configured precision.
|
|
|
|
For a precision threshold of `c`, the implementation that we are using requires about `c * 8` bytes.
|
|
|
|
The following chart shows how the error varies before and after the threshold:
|
|
|
|
|
|
|
|
For all 3 thresholds, counts have been accurate up to the configured threshold. Although not guaranteed,
|
|
this is likely to be the case. Accuracy in practice depends on the dataset in question. In general,
|
|
most datasets show consistently good accuracy. Also note that even with a threshold as low as 100,
|
|
the error remains very low (1-6% as seen in the above graph) even when counting millions of items.
|
|
|
|
The HyperLogLog++ algorithm depends on the leading zeros of hashed values, the exact distributions of
|
|
hashes in a dataset can affect the accuracy of the cardinality.
|