Set Headers directive

The SET-HEADERS directive sets the names of columns, in the order they are specified.

Syntax

set-headers :columm [,:column]*

The column specifies the new name of an existing column or columns.

Usage Notes

The most common use of the SET-HEADERS directive is to set the name of columns when a CSV file is parsed. The column names will be applied to the record starting from the first field, in the order that they are specified.

Examples

Using this record as an example:

{ "body": "1,2,3,4,5" }

If you have parsed this body using the PARSE-AS-CSV directive:

parse-as-csv :body , false

the resulting record would be:

If you then apply the SET COLUMNS directive:

This would generate a record that has these column names:

Note that the last field (body_5) was not assigned the expected name.

In order to correct this, either rename all the columns using:

resulting in this record:

or use a DROP directive:

The result would then be this record:

Common Mistakes

When using the SET-HEADERS directive, the number of fields in the record should be same as number of column names in the SET-HEADERS directive. If they are not, then this directive will partially name the record fields.

The names of the columns are in a single option, separated by commas. Separating by spaces will set only the name of the first column.

When this directive is executed in a pipeline and the field "Name of field" to be transformed is set to *, then all fields are added to the record causing issues with the naming of the columns, as it would also include column names that are coming from the input.

Created in 2020 by Google Inc.