Unpivot Transform

NOTE: Transforms are a part of the underlying language that is not directly accessible to users. This content is maintained for reference purposes only.

Reshapes the data by merging one or more columns into key and value columns. Keys are the names of input columns, and value columns are the cell values from the source.

Rows of data are duplicated, once for each input column.

The unpivot column can be applied to multiple columns in the same transform. All columns are un-pivoted into the same key and value columns. When this transform is applied to two columns, the number of rows in the dataset is doubled.

This transform is the opposite of the pivot transform, which converts a set of column values into distinct columns. See Pivot Transform.

Basic Usage

Single- or multi-column example:

You can specify single columns or comma-separated sets of columns.

unpivot col: FirstName, MiddleInitial

Output: Converts the values in the columns FirstName and MiddleInitial into separate key and value columns.

Column range example:

You can also specify ranges of columns using the tilde (~) operator:

unpivot col:Column1~Column20

Output: Converts all of the values in columns between Column1 and Column20 into key and value columns.

Parameters

unpivot col: column_ref [groupEvery: int_num]

TokenRequired?Data TypeDescription
unpivotYtransformName of the transform
colYstringName of source column or columns
groupEveryNstringIf specified, this parameter defines the number of individual key-value pairs to store in each generated column. Default is 1.

For more information on syntax standards, see Language Documentation Syntax Notes.

col

Identifies the column or columns to which to apply the transform. You can specify one or more columns.

To specify multiple columns:

  • Discrete column names are comma-separated. Values for column names are case-sensitive.
  • Where applicable, a range of values can be specified using a tilde (~).

Usage Notes:

Required?Data Type
YesString (column name)

groupEvery

Specifies the number of output key-value pair columns to produce after unpivoting.

This optional parameter is used to create multiple sets of key-value pair columns in the output. The columns listed in the col parameter are placed into each pair of output key-value columns sequentially. After all key-value pair columns are filled in a record, the next column is placed into the first key-value pair column of the next record.

By default, this value is 1, meaning that each column specified in the transform is rendered into a new record in a single pair of key-value columns.

Usage Notes:

Required?Data Type
NoInteger (positive)

Examples

Example - Basic Unpivot

Source:

productNameproductColorproductSize
WhizbangredM
Whizbangred, blueL
WhizbanggreenM
BangwhizredS
BangwhizblueM
BangwhizredS

Tranform:

After you have created a header, if necessary, add the following transform:

unpivot col:productColor

Results:

productNameproductSizekeyvalue
WhizbangMproductColorred
WhizbangLproductColorred, blue
WhizbangMproductColorgreen
BangwhizSproductColorred
BangwhizMproductColorblue
BangwhizSproductColorred

Extended:

Note how each instance of a value results in a separate row; duplicate values are included. For a single-column unpivot, this transform results in the same number of rows as the source.

  • Since the value is treated as a string, the value red, blue is treated as one value.

Now, edit the transform you just added. Replace it with the following, which includes the productSize key as part of the transform:

unpivot col:productColor,productSize

Results:

productNamekeyvalue
WhizbangproductColorred
WhizbangproductSizeM
WhizbangproductColorred, blue
WhizbangproductSizeL
WhizbangproductColorgreen
WhizbangproductSizeM
BangwhizproductColorred
BangwhizproductSizeS
BangwhizproductColorblue
BangwhizproductSizeM
BangwhizproductColorred
BangwhizproductSizeS

Row keys alternate based on the order in which the source columns are specified in the transform. Since the transform specifies two columns, the number of key-value pairs is doubled, which results in a dataset that has twice as many rows as the source.

Example - Basic Pivot with groupEvery

From the previous example, modify the unpivot transform to be the following:

unpivot col:productColor,productSize groupEvery:2

Results:

productNamekey1value1key2value2
WhizbangproductColorredproductSizeM
WhizbangproductColorred, blueproductSizeL
WhizbangproductColorgreenproductSizeM
BangwhizproductColorredproductSizeS
BangwhizproductColorblueproductSizeM
BangwhizproductColorredproductSizeS

Was this page helpful? Let us know how we did:

Send feedback about...

Google Cloud Dataprep Documentation
Need help? Visit our support page.