The ML.HOLIDAY_INFO function
This document describes theML.HOLIDAY_INFO function, which you can use toreturn the list of holidays being modeled by anARIMA_PLUSorARIMA_PLUS_XREGtime series forecasting model.
Syntax
ML.HOLIDAY_INFO( MODEL `PROJECT_ID.DATASET.MODEL_NAME`)
Arguments
ML.HOLIDAY_INFO takes the following arguments:
PROJECT_ID: your project ID.DATASET: the BigQuery dataset that containsthe model.MODEL_NAME: the name of the model.
Output
ML.HOLIDAY_INFO returns the following columns:
region: aSTRINGvalue that identifies the holiday region.holiday_name: aSTRINGvalue that identifies the holiday.primary_date: aDATEvalue that identifies the date the holidayfalls on.preholiday_days: anINT64value that identifies the start of theholiday window around the holiday that was taken into account whenmodeling.postholiday_days: anINT64value that identifies the end of theholiday window around the holiday that was taken into account whenmodeling.
Example
The following example returns the results for a model that uses acustom holiday:
SELECT*FROMML.HOLIDAY_INFO(MODEL`mydataset.arima_model`);
The output looks similar to the following:
+-----------------------+--------------+-----------------+------------------+| region | holiday_name | primary_date | preholiday_days | postholiday_days |+--------------------------------------------------------+------------------+| US | Members day | 2001-10-21 | 3 | 1 |+-----------------------+--------------+-----------------+------------------+| US | Members day | 2002-10-22 | 3 | 1 |+-----------------------+--------------+-----------------+------------------+| US | Members day | 2003-10-21 | 3 | 1 |+-----------------------+--------------+-----------------+------------------+| US | Members day | 2004-10-23 | 3 | 1 |+-----------------------+--------------+-----------------+------------------+
Limitation
- Results returned by
ML.HOLIDAY_INFOonly indicate the holiday informationused during model fitting. They don't necessarily indicate the detection of aholiday effect. UseML.EXPLAIN_FORECASTinstead for actual holiday effect results.
What's next
- For more information about model evaluation, seeBigQuery ML model evaluation overview.
- For more information about supported SQL statements and functions for timeseries forecasting models, seeEnd-to-end user journeys for time series forecasting models.
Except as otherwise noted, the content of this page is licensed under theCreative Commons Attribution 4.0 License, and code samples are licensed under theApache 2.0 License. For details, see theGoogle Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.
Last updated 2025-12-15 UTC.