Skip to main content

Select

Choose, rename, reorder, and cast columns.

Sockets

SocketDirectionDescription
inputInputData to transform
outputOutputData with selected columns

Configuration

The Select tool shows a table of all available columns with the following controls:

ColumnDescription
CheckboxInclude/exclude the column from output
NameOriginal column name (read-only)
RenameNew name for the column (optional)
TypeTarget data type (optional, for casting)
SizeFor string types, maximum character length (optional)

Column Order

Drag rows to reorder columns. The output will have columns in the order shown in the configuration.

Operations

Selecting Columns

Check the columns you want to keep, uncheck those you want to remove.

At least one column must be selected.

Renaming Columns

Enter a new name in the "Rename" field. Leave empty to keep the original name.

Reordering Columns

Drag and drop rows to change column order. The first row becomes the first column, etc.

Casting Types

Select a target type from the dropdown to convert the column. Available types:

CategoryTypes
IntegerInt8, Int16, Int32, Int64, UInt8, UInt16, UInt32, UInt64
FloatFloat32, Float64
StringString, Utf8
BooleanBoolean
TemporalDate, Datetime, Time, Duration
OtherCategorical

Truncating Strings

For String/Utf8 columns, enter a size limit to truncate values to that many characters.

Examples

Keep Only Specific Columns

  1. Uncheck all columns you don't need
  2. Keep checked only the columns you want

Rename a Column

  1. Find the column in the list
  2. Enter the new name in the "Rename" field

Cast String to Integer

  1. Find the column containing numeric strings
  2. Select "Int64" from the Type dropdown
  3. Ensure the data is actually numeric, or the cast will fail

Reorder Columns

  1. Drag the "id" row to the top
  2. Drag "name" below it
  3. Arrange remaining columns as needed

Notes

  • Empty configuration: If no columns are configured, all columns pass through unchanged
  • Duplicate names: An error is shown if you rename two columns to the same name
  • Cast failures: If a value can't be cast to the target type, execution will fail
  • Row count unchanged: Select only affects columns, not rows