---
title: "init(dictionaries:)"
framework: createml
role: symbol
role_heading: Initializer
path: "createml/mluntypedcolumn/init(dictionaries:)"
---

# init(dictionaries:)

Creates a new column of machine learning dictionaries by converting the elements of another column.

## Declaration

```swift
init(dictionaries: MLUntypedColumn)
```

## Parameters

- `dictionaries`: A column with elements that are convertible to Create ML dictionaries.

## Return Value

Return Value A new untyped column of doubles; otherwise an invalid column if any element of the given column cannot be converted to MLDataValue.DictionaryType.

## Discussion

Discussion Use this initializer to create a column of dictionaries from another column. As an example, to create a column with this initializer, first start with a column that is convertible to dictionaries. let intDictionaryString = "{1:\"one\", 2:\"two\", 3:\"three\"}" let intDictionaryString2 = "{4:\"four\", 5:\"five\", 6:\"six\"}" let stringsColumn = MLUntypedColumn([intDictionaryString, intDictionaryString2])

print(stringsColumn) /* Prints...  ValueType: String Values: [{1:"one", 2:"two", 3:"three"}, {4:"four", 5:"five", 6:"six"}]  */ Then use init(dictionaries:) to convert the column to a column of dictionaries. let dictionaryColumn = MLUntypedColumn(dictionaries: stringsColumn) print(dictionaryColumn) /* Prints...  ValueType: Dictionary  Values:        [[1: one, 3: three, 2: two], [4: four, 5: five, 6: six]]  */

## See Also

### Creating an untyped column by converting another column

- [init(ints:)](createml/mluntypedcolumn/init(ints:).md)
- [init(doubles:)](createml/mluntypedcolumn/init(doubles:).md)
- [init(strings:)](createml/mluntypedcolumn/init(strings:).md)
- [init(sequences:)](createml/mluntypedcolumn/init(sequences:).md)
- [init(multiArrays:)](createml/mluntypedcolumn/init(multiarrays:).md)