Some Wrangles require authentication to the WrangleWorks servers. See the installation instructions
Once data is read from a source, Wrangles can be used to transform the data. Some Wrangles apply simple transformations, while others use sophisticated machine learning models to infer information from the data.
Wrangles are applied sequentially, starting from the base state of the source. Each Wrangle is applied to the whole table.
Each Wrangle typically has an input and an output. An input can be the name of one column, or a list of multiple columns. If no output is provided, the input column will be overwritten, otherwise a new column will be created.
read:
# Generate some example data
- test:
rows: 3
values:
column1: an example sentence
wrangles:
# If output is omitted, the input will be overwritten
- convert.case:
input: column1
case: upper
# If output is specified, a new column will be created
- split.tokenize:
input: column1
output: column2
| → |
| → |
|