Icon

02_​Model_​Deployment_​SQL_​Java_​recoding

Model Deployment SQL Recoding
Output - Static to file or database - Dynamic through REST Interface - Static visual: report or email - Dynamic visual: dashboard - External application update Input - Static data from file or database - Dynamic data through REST Interface Deployment Mode - Remote execution from server - Scheduled execution from server - External application - REST API from Server - Guided Execution from WebPortal Apply Model - PMML predictor- Dedicated predictor- SQL model recoding- Executable JAR model Model Deployment with SQL Recoding (external application) This workflow: - reads new unseen data from database (.sqlite format), - reads the trained model (.pmml format), - generates SQL query from PMML model, - applies the model by executing the database query.Blog post available at https://www.knime.org/blog/seven_modes_deploymentVideo available at https://youtu.be/Og7VZOJhsOc Fix "flights" table (if the workflow was previously executed) Overwrite "flights" tableGenerate SQL queryRead the modelCreate table"predictions"Select "flights" tableUpdate input database query(apply the model)Append column to "flights",copy values from "predictions" to"flights"Connect to SQLitedatabasetranslate PMML modeinto Java bytecode write JARflights originaltableConnect to SQLitedatabase Database Writer PMML to SQL PMML Reader Database ConnectionTable Writer Database TableSelector Database Query Database SQLExecutor SQLite Connector PMML Compiler Model Writer Table Reader SQLite Connector Output - Static to file or database - Dynamic through REST Interface - Static visual: report or email - Dynamic visual: dashboard - External application update Input - Static data from file or database - Dynamic data through REST Interface Deployment Mode - Remote execution from server - Scheduled execution from server - External application - REST API from Server - Guided Execution from WebPortal Apply Model - PMML predictor- Dedicated predictor- SQL model recoding- Executable JAR model Model Deployment with SQL Recoding (external application) This workflow: - reads new unseen data from database (.sqlite format), - reads the trained model (.pmml format), - generates SQL query from PMML model, - applies the model by executing the database query.Blog post available at https://www.knime.org/blog/seven_modes_deploymentVideo available at https://youtu.be/Og7VZOJhsOc Fix "flights" table (if the workflow was previously executed) Overwrite "flights" tableGenerate SQL queryRead the modelCreate table"predictions"Select "flights" tableUpdate input database query(apply the model)Append column to "flights",copy values from "predictions" to"flights"Connect to SQLitedatabasetranslate PMML modeinto Java bytecode write JARflights originaltableConnect to SQLitedatabase Database Writer PMML to SQL PMML Reader Database ConnectionTable Writer Database TableSelector Database Query Database SQLExecutor SQLite Connector PMML Compiler Model Writer Table Reader SQLite Connector

Nodes

Extensions

Links