This document describes the ML.ONE_HOT_ENCODER function, which lets you encode a string expression using a one-hot or dummy encoding scheme.

The encoding vocabulary is sorted alphabetically. NULL values and categories that aren't in the vocabulary are encoded with an index value of 0. If you use dummy encoding, the dropped category is encoded with a value of 0.

When used in the TRANSFORM clause, the vocabulary and dropped category values calculated during training, along with the top k and frequency threshold values that you specified, are automatically used in prediction.


ML.ONE_HOT_ENCODER(string_expression [, drop] [, top_k] [, frequency_threshold]) OVER()


ML.ONE_HOT_ENCODER takes the following arguments:

  • string_expression: the STRING expression to encode.
  • drop: a STRING value that specifies whether the function drops a category. Valid values are as follows:
    • none: Retain all categories. This is the default value.
    • most_frequent: Drop the most frequent category found in the string expression. Selecting this value causes the function to use dummy encoding.
  • top_k: an INT64 value that specifies the number of categories included in the encoding vocabulary. The function selects the top_k most frequent categories in the data and uses those; categories below this threshold are encoded to 0. This value must be less than 1,000,000 to avoid problems due to high dimensionality. The default value is 32,000.
  • frequency_threshold: an INT64 value that limits the categories included in the encoding vocabulary based on category frequency. The function uses categories whose frequency is greater than or equal to frequency_threshold; categories below this threshold are encoded to 0. The default value is 5.


ML.ONE_HOT_ENCODER returns an array of struct values, in the form ARRAY<STRUCT<INT64, FLOAT64>>. The first element in the struct provides the index of the encoded string expression, and the second element provides the value of the encoded string expression.


The following example performs dummy encoding on a set of string expressions. It limits the encoding vocabulary to the ten categories that occur the most frequently in the data and that also occur zero or more times.

SELECT f, ML.ONE_HOT_ENCODER(f, 'most_frequent', 10, 0) OVER () AS output
FROM UNNEST([NULL, 'a', 'b', 'b', 'c', 'c', 'c', 'd', 'd']) AS f

The output looks similar to the following:

|  f   | output.index | output.value |
| NULL |  0           |  1.0         |
| a    |  1           |  1.0         |
| b    |  2           |  1.0         |
| b    |  2           |  1.0         |
| c    |  3           |  0.0         |
| c    |  3           |  0.0         |
| c    |  3           |  0.0         |
| d    |  4           |  1.0         |
| d    |  4           |  1.0         |

What's next