Purpose
Creates one row per combination of key fields and calculates a sum, count, minimum, maximum or average. Without keys, it calculates a total for the entire dataset; keys alone, without measures, produce unique combinations.
Inputs and outputs
One input, in. The out output contains the keys and specified aggregate fields.
Parameters
| Parameter | Settings and values | Options and limits |
|---|---|---|
excluded | Fields excluded from grouping roles; they do not appear in the result. | Count: 0–16384 |
keys | Fields whose combined values define a group. | Count: 0–100 |
aggregates | Measures: column — the field, op — the operation, name — the result name. | Count: 0–100 |
decimal | The decimal separator for numeric values in the input data. | Options: ., ,Default: . |
Aggregate operations
| Parameter | What it configures |
|---|---|
| sum | The sum of all normalised values. |
| count | The number of rows in the group, including empty values. |
| min / max | The smallest or largest normalised number. |
| avg | The sum of normalised values divided by the number of rows in the group. |
Step-by-step configuration
- Define the result’s granularity and select keys, such as Product and Month.
- Add measures and choose sum/count/min/max/avg. count does not require a value field.
- Set unique result names and the decimal separator.
- Check the group count and total against a small, known dataset.
Rules and diagnostics
- Empty values and ordinary text count as zero in numeric aggregates. Therefore, avg includes those rows in its denominator.
- Invalid numeric representations, cell errors and unsupported types retain their diagnostics and may stop the block.
- A measure field does not automatically remain a key/excluded field at the same time; grouping normalises its role.
- Key types matter: do not convert a text code such as “001” to a number unnecessarily.
Configuration example
Group sales by Product and Month, using Value → sum named Total, and add count named Rows. For A/Jan/10 and A/Jan/15, the result is A/Jan/25/2.