Group and GroupBy have similar names. However, they serve different use cases.
|
|
Group |
GroupBy |
|
Used with functions |
asPercent |
Pairs |
| Output | Same metrics | Grouped metrics |
SAMPLE USE CASE FOR GROUP
| Metrics | A large number of metrics representing bounce_rate measure with dimensions device_category, OS, country |
| Challenge | To understand the percentiles of bounce_rate. Using asPercent directly on a large number of metrics will not help. |
| Solution | Use Group based on OS and then activate asPercent over that grouping. |
| Result | Get the percentiles based on OS |
SAMPLE USE CASE FOR GROUPBY
| Metrics | A large number of metrics representing bounces measure with dimensions device_category, OS, country |
| Challenge | To understand the ratio of bounces for a specific OS in a specific country compared to the entire amount of bounces in that country |
| Solution |
Use Pairs (or ratioPairs) In the numerator use GroupBy based on OS and country In the Denominator use GroupBy based on country |
| Result | Get the ratio based on OS for each country |
For more details about the Group and GroupBy metric functions, see