Grouping behavior in Pliable determines how records are processed, combined, or aggregated during data preparation. These options allow for fine-tuned control over how data is treated based on business or analysis requirements.

How to Configure Grouping Behavior

Start a new cleaning step by clicking on “Start Cleaning.” from the data source node screen, or navigate to an existing cleaning step.
Locate the "Grouping Behavior" dropdown menu.

Select the desired grouping option:
- No Grouping: Leaves records as-is.
- Merge: Choose the columns to define a match and merge.
- Summarize: Specify the grouping criteria and select aggregation functions.
- Identify: Specify the columns to check for common attributes.
Preview the output to ensure the grouping behavior meets your requirements.

Grouping Behaviors

Grouping Option	Description
No Grouping	Processes records individually without merging or aggregating. Each record is treated as a unique row in the output.
Merge	Combines matching records from input data into a single record. Useful for deduplication or consolidating datasets.
Summarize	Aggregates records based on shared criteria (e.g., grouping by category and summing values). Ideal for generating roll-ups or summaries.
Identify	Detects (but does not merge) records that share common attributes, marking them for further processing.

Detailed Behavior

1. No Grouping

Use Case: When each record needs to remain distinct and no grouping or merging logic is required.
Example: Exporting transactional data where every row represents a single event or activity.

2. Merge

Use Case: Consolidates duplicate records or records that share a key column. Helps clean data by eliminating redundancies.
Logic:
- Matches records based on specified criteria (e.g., primary key or unique identifier).
- Combines the data into a single record.
Example: Merging customer records with the same email address.

3. Summarize

Use Case: Groups records based on one or more columns and applies aggregation methods (e.g., SUM, AVG, MAX).
Logic:
- Records are grouped by the specified key(s).
- Metrics are aggregated based on selected columns and functions.