Home - News ( United States | United Kingdom | Italy | Germany ) - Football scores

Showing content from https://cloud.google.com/spanner/docs/reference/standard-sql/ml-functions below:

Machine learning functions in GoogleSQL | Spanner

Stay organized with collections Save and categorize content based on your preferences.

GoogleSQL for Spanner supports the following machine learning (ML) functions.

ML.PREDICT(input_model, input_relation[, model_parameters])

input_model:
  MODEL model_name

input_relation:
  { input_table | input_subquery }

input_table:
  TABLE table_name

model_parameters:
  STRUCT(parameter_value AS parameter_name[, ...])

Description

ML.PREDICT is a table-valued function that helps to access registered machine learning (ML) models and use them to generate ML predictions. This function applies ML computations defined by a model to each row of an input relation, and then, it returns the results of the predictions.

Supported Argument Types

• input_model: The model to use for predictions. Replace model_name with the name of the model. To create a model, see CREATE_MODEL.
• input_relation: A table or subquery upon which to apply ML computations. The set of columns of the input relation must include all input columns of the input model; otherwise, the input won't have enough data to generate predictions and the query won't compile. Additionally, the set can also include arbitrary pass-through columns that will be included in the output. The order of the columns in the input relation doesn't matter. The columns of the input relation and model must be coercible.
• input_table: The table containing the input data for predictions, for example, a set of features. Replace table_name with the name of the table.
• input_subquery: The subquery that's used to generate the prediction input data.
• model_parameters: A STRUCT value that contains parameters supported by model_name. These parameters are passed to the model inference.

Return Type

A table with the following columns:

• Model outputs
• Pass-through columns from the input relation

Examples

The examples in this section reference a model called DiamondAppraise and an input table called Diamonds with the following columns:

• DiamondAppraise model:

  Input columns Output columns value FLOAT64 value FLOAT64 carat FLOAT64 lower_bound FLOAT64 cut STRING upper_bound FLOAT64 color STRING(1)

• Diamonds table:

  Columns Id INT64 Carat FLOAT64 Cut STRING Color STRING

The following query predicts the value of a diamond based on the diamond's carat, cut, and color.

SELECT id, color, value
FROM ML.PREDICT(MODEL DiamondAppraise, TABLE Diamonds);

+----+-------+-------+
| id | color | value |
+----+-------+-------+
| 1  | I     | 280   |
| 2  | G     | 447   |
+----+-------+-------+

You can include model-specific parameters. For example, in the following query, the maxOutputTokens parameter specifies that content, the model inference, can contain 10 or fewer tokens. This query succeeds because the model TextBison contains a parameter called maxOutputTokens.

SELECT prompt, content
FROM ML.PREDICT(
  MODEL TextBison,
  (SELECT "Is 13 prime?" as prompt), STRUCT(10 AS maxOutputTokens));

+----------------+---------------------+
| prompt         | content             |
+----------------+---------------------+
| "Is 13 prime?" | "Yes, 13 is prime." |
+----------------+---------------------+

You can use ML.PREDICT in any DQL/DML statements, such as INSERT or UPDATE. For example:

INSERT INTO AppraisedDiamond (id, color, carat, value)
SELECT
  1 AS id,
  color,
  carat,
  value
FROM
  ML.PREDICT(MODEL DiamondAppraise,
  (
    SELECT
      @carat AS carat,
      @cut AS cut,
      @color AS color
  ));

Except as otherwise noted, the content of this page is licensed under the Creative Commons Attribution 4.0 License, and code samples are licensed under the Apache 2.0 License. For details, see the Google Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.

Last updated 2025-08-07 UTC.

[[["Easy to understand","easyToUnderstand","thumb-up"],["Solved my problem","solvedMyProblem","thumb-up"],["Other","otherUp","thumb-up"]],[["Hard to understand","hardToUnderstand","thumb-down"],["Incorrect information or sample code","incorrectInformationOrSampleCode","thumb-down"],["Missing the information/samples I need","missingTheInformationSamplesINeed","thumb-down"],["Other","otherDown","thumb-down"]],["Last updated 2025-08-07 UTC."],[],[]]

RetroSearch is an open source project built by @garambo | Open a GitHub Issue

Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo

HTML: 3.2 | Encoding: UTF-8 | Version: 0.7.4