Hello,
I’m testing InfluxDB and studying possible constraints. I’ve noticed that the exact cardinality measurement is greater than the current number of possibilities of tag combinations. The estimated measure seems to follow well the number of possible series.
Is this exact measurement correct? If so, how are series being defined if not by the tag values present in the data?
To illustrate, I have 3 tags, ‘tag_A’ has 10k possible values, ‘tag_B’ has 2 possible values and ‘tag_C’ has only one value, yielding 20k possible combinations of tag values.
I have currently added 50k data points to my measurement. Results are as follows:
> show series cardinality
cardinality estimation
----------------------------
18413
> show series exact cardinality
name: test_measurement
count
-------
48842
