Pivot Data

A pivot table summaries data that is sourced from another table. Using pivot tables, you can calculate aggregating functions, such as sums, maximums, and averages for one or more columns of data. Optionally, these sums can be performed across groups of values from one column and broken out in columns based on the values in another. In Designer Cloud Enterprise Edition, a pivot table is composed of the following basic elements:

Building a Pivot Table

Pivot tables are very powerful tools for summarizing and visualizing large-scale volumes of data. In  Designer Cloud Enterprise Edition, search for pivot table in the Search panel to create one.

In your flows, you may find it useful to create your pivot tables in independent recipes that are chained from your primary recipe. For more information, see Create Branching Outputs.

Example Data

Pivot tables are perhaps best explained by example. The following table snippet captures transactional data from a number of stores for a range of products across a set of dates. Transactional values include total sales, quantity, and cost (POS_Sales, POS_Qty, and POS_Cost):

Available Aggregations

The Pivot data transformation supports use of any aggregation function. For more information, see Aggregate Functions.

Simple Pivot Table

From the above, suppose you are interested in the sales from each store for each product. You can use the following transformation to compute these aggregated calculations:

In the above transformation:

• The Column labels entry specifies the column whose values make up the calculated columns of the pivot table. The calculation is performed across each of these values. In this case, each column contains calculations for separate store numbers.
• The Row labels entry specifies the column whose values define the grouping of the calculations. In this case, the sum of the sales column is performed for each product description value for each store.
• The Values entry specifies the aggregation function to compute for each cell in the new table. In this case, you are generating the sum of sales for each product description in each store.

• By default, this transformation generates a maximum of 50 new columns. However, if the column used for your Column labels contains more than 50 values, you may want to raise this value.

  NOTE: Avoid creating datasets wider than 2500 columns. Very wide datasets can cause performance degradation.

Results:

Conditional Aggregations

Suppose you are interested in only in the sum of sales for store numbers 1-3. To capture a more limited dataset, you can use the SUMIF aggregation function:

Most aggregation functions have a conditional (*IF) variant. See Aggregate Functions.

Multiple Aggregation Levels

None of the axes of a pivot table is limited to a single dimension. You can have multiple Column labels, Row labels, and Values (formulas). In the following transformation, aggregations have been further broken out by date, and an additional formula (Value) has been added.

Results:

Group By

If you wish to maintain the original dataset values, you can perform aggregation calculations within a single column. For more information, see Create Aggregations.

Values to Columns

Similar to pivot, the Convert values to columns transformation converts individual values within a column to independent columns in the dataset. For each row, if the value represented by the column is present in the original data, one value is added (e.g. Yes). If it's missing, another value is inserted (e.g. No).

In the following, the values in the Store_Nbr column have been converted to individual columns:

In the above:

• Fill when present identifies the string literal value to insert if the row contains the column's value (Yes).
• Fill when missing identifies the string literal value to insert if the row does not contain the column's value (empty).

• Max number of columns to create places a limit on the total number of columns that the application is permitted to create. In this case, the limit is set to 250 since the known number of stores is 250.

  Tip: It's a good habit to set limits on the maximum number of columns to create. Data can become sparse or unwieldy if limits are not considered.

Results: