SELECT statement

Description

The SELECT statement is used to get a predictions from the model table. The data is not persistent and is returned on the fly as a result-set. The basic syntax for selecting from the model is:

Syntax

SELECT [target_variable], [target_variable]_explain
FROM mindsdb.[predictor_name]
WHERE [column]=[value] 
    AND [column]=[value];

Grammar matters

Ensure that there are no spaces between the column name, equal sign and value. Ensure to not use any quotations for numerical values and singular quotes for strings

On execution, you should get:

+----------+----------+-------------------+-----------------------------------------------------------------------------------------------------------------------------------------------+
| [column] | [column] | [target_variable] | [target_variable]_explain                                                                                                                     |
+----------+----------+-------------------+-----------------------------------------------------------------------------------------------------------------------------------------------+
| [value]  | [value]  | [predicted_value] | {"predicted_value": 4394, "confidence": 0.99, "anomaly": null, "truth": null, "confidence_lower_bound": 4313, "confidence_upper_bound": 4475} |
+----------+----------+-------------------+-----------------------------------------------------------------------------------------------------------------------------------------------+

Where:

Expressions	Description
[target_variable]	Name of the column to be predicted
[target_variable]_explain	JSON object that contains additional information as confidence_lower_bound, confidence_upper_bound, anomaly, truth.
[predictor_name]	Name of the model to be used to make the prediction
WHERE [column]=[value] AND ...	WHERE clause used to pass the input data to make the prediction

Example

The following SQL statement selects a rental_price prediction from the home_rentals_model for a property that has the attributes named after the WHERE expression:

SELECT location, neighborhood, days_on_market, rental_price, rental_price_explain
FROM mindsdb.home_rentals_model1
    WHERE sqft = 823
    AND location='good'
    AND neighborhood='downtown'
    AND days_on_market=10;

On execution,

+----------+--------------+----------------+--------------+-----------------------------------------------------------------------------------------------------------------------------------------------+
| location | neighborhood | days_on_market | rental_price | rental_price_explain                                                                                                                          |
+----------+--------------+----------------+--------------+-----------------------------------------------------------------------------------------------------------------------------------------------+
| good     | downtown     | 10             | 4394         | {"predicted_value": 4394, "confidence": 0.99, "anomaly": null, "truth": null, "confidence_lower_bound": 4313, "confidence_upper_bound": 4475} |
+----------+--------------+----------------+--------------+-----------------------------------------------------------------------------------------------------------------------------------------------+

Bulk predictions

You can also make bulk predictions by joining a table with your model:

SELECT t.rental_price as real_price,
    m.rental_price as predicted_price,
    t.sqft, t.location, t.days_on_market
FROM example_db.demo_data.home_rentals as t
JOIN mindsdb.home_rentals_model as m limit 100