Pivot¶
The Pivot processor transforms multiple rows into columns.
Example¶
Let’s imagine we are processing financial data, which is in ‘multiple lines’ format
Input¶
| Company | Type | Value |
|---|---|---|
| Comp.A | Revenue | 42M |
| Comp.A | Raw Margin | 9M |
| Comp.B | Revenue | 137M |
| Comp.B | Raw Margin | 3M |
| Comp.B | Net income | -11M |
Process¶
If we pivot with:
- Index column: Company
- Labels column: Type
- Values column: Value
We obtain:
| Company | Revenue | Raw Margin | Net income |
|---|---|---|---|
| Comp.A | 42M | 9M | |
| Comp.B | 137M | 3M | -11M |
Details¶
Pivot generates a new row for each change of value of the index column. The generated rows will contain one column for each value of the labels column. The output values will be the values of the values column.
Important note: this processor assumes that the values of the index column are sorted. IE:
- OK
| idx1 | label1 | v1 |
| idx1 | label2 | v2 |
| idx2 | label1 | v3 |
- NOK
| idx1 | label1 | v1 |
| idx2 | label1 | v3 |
| idx1 | label2 | v2 |
In the latter case, 3 rows will be generated.
Also, if several rows have the same (index,label), then only the value corresponding to the last one is kept.