Page tree

Outdated release! Latest docs are Release 8.7: Header Transform


NOTE: Transforms are a part of the underlying language, which is not directly accessible to users. This content is maintained for reference purposes only. For more information on the user-accessible equivalent to transforms, see Transformation Reference.

Uses one row from the dataset sample as the header row for the table. Each value in this row becomes the name of the column in which it is located. 

This transform might be automatically applied as one of the first steps of your recipe. See Initial Parsing Steps.

NOTE: If source row number information is not available due to changes in the dataset, this transform may not be available.

Basic Usage

header sourcerownumber: 4

Output: The values from Row #4 of the original dataset are used, if available, as the names for each column. If the row is not available, the specified row data can be retrieved, and the transform fails.

Syntax and Parameters

header sourcerownumber: row_num

TokenRequired?Data TypeDescription
headerYtransformName of the transform
sourcerownumberYinteger (positive)Row number from the original data to use as the header.

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


The sourcerownumber parameter defines the row number to apply to the transform step. 

This parameter references the original row number of the sample in the dataset.

  • sourceownumber parameter must be an integer that is less than or equal to the total number of rows in the original sample.
  • If the corresponding row has been deleted from the dataset, the transform step generates an error.



header sourcerownumber: 4

Output: Uses row #4 from the source row numbers of the sample as the header the columns. 

Usage Notes:

Required?Data Type
Yesinteger (positive)



Tip: For additional examples, see Common Tasks.

Example - Header from row that is not the first one


You have imported the following racer data on heat times from a CSV file. When loaded in the Transformer page, it looks like the following:

1RacerHeat 1Heat 2Heat 3
2Racer X37.2238.2237.61
3Racer Y41.33DQ38.04
4Racer Z39.2739.0438.85

In the above, the (rowId) column references the row numbers displayed in the data grid; it is not part of the dataset. This information is available when you hover over the black dot on the left side of the screen.


You have examined the best performance in each heat according to the sample. You then notice that the data contains headers, but you forget how it was originally sorted. The data now looks like the following:

1Racer Y41.33DQ38.04
2RacerHeat 1Heat 2Heat 3
3Racer X37.2238.2237.61
4Racer Z39.2739.0438.85

While you can undo your sort steps to return to the original sort order, this approach works best if you did not include other steps in between that are based on the sort order.

If you have steps that require retaining your sort steps, you can revert to the original sort order by adding this transform step:

NOTE: Source row information may be lost after operations such as joins, unions, and aggregations are performed. In these cases, you cannot sort by the source row information. You may be able to generate a column of source row number earlier in your recipe.


Then, you can create the header with the following simple step:

header sourcerownumber:1


After you have applied the last header transform, your data should look like the following:

3Racer Y41.33DQ38.04
2Racer X37.2238.2237.61
4Racer Z39.2739.0438.85

You can sort by the Racer column in ascending order to return to the original sort order.

This page has no comments.