DataAI ETL InterSystems IRIS - Customer-Ready Evaluation

Product: InterSystems IRIS

Distribution: Customer-ready evaluation

Source: DataAI_ETL_IRIS_Evaluation/docs/INSTALLATION_AND_USAGE.md inside DataAI_ETL_IRIS_Evaluation.zip

Package SHA-256: f30ff347ab2873d19726e4c3ac0f36d743a66eab9dbaba248b5103e15120725a

Download customer-ready package | Download checksum | Read the evaluation license

Start here: Download one customer-ready ZIP. It already contains the DataAI libraries, product assets, documentation, samples, licenses, internal checksums, and local verification files for this integration. Spark, Hadoop, vendor products, and proprietary drivers remain customer-provided.

InterSystems IRIS: Customer Evaluation Installation and Usage

This is the customer-ready DataAI ETL evaluation package from Yanbor LLC, provider of the DataAI product. DataAI runs inside the customer's pipeline; it is not a hosted service and makes no required DataAI network call.

Prerequisites

Installation and first verification

  1. Review LICENSE.md and use an isolated IRIS evaluation namespace.
  2. Run smoke-test\run-smoke-test.ps1 for the local DataAI library check.
  3. Execute smoke-test\iris-live\setup-evaluation.sql in the isolated namespace.
  4. Run smoke-test\iris-live\run-smoke-test.ps1 with the path to the customer-supplied IRIS JDBC driver.
  5. Require DATAAI_IRIS_SMOKE_TEST_PASS and inspect the written DataAI tables with verify-results.sql.

Evaluation boundary

Package layout

Local DataAI Library Smoke Test

This test uses only fictional in-memory rows. It starts local Spark, executes normalization, profiling, required and numeric quality rules, clean/rejected routing, the complete function-catalog check, and iterative matrix balancing.

Requirements

  • Java 17.
  • Apache Spark 3.5.0 with Scala 2.12.
  • spark-submit available on PATH, or supplied with -SparkSubmit.

Spark and Hadoop are customer-provided runtimes and are intentionally not bundled in the DataAI evaluation ZIP.

Run

From the extracted package root:

.\smoke-test\run-smoke-test.ps1

If the customer's Windows policy blocks direct script invocation and the customer's security policy permits a process-only override, run:

powershell.exe -NoProfile -ExecutionPolicy Bypass -File `
  .\smoke-test\run-smoke-test.ps1

This option applies only to that PowerShell process; it does not change the computer's execution policy.

For an explicit executable path:

.\smoke-test\run-smoke-test.ps1 `
  -SparkSubmit 'C:\spark\bin\spark-submit.cmd'

Success requires exit code 0 and a DATAAI_*_SMOKE_TEST_PASS marker. This local check validates the delivered DataAI libraries. It does not replace the documented validation on Databricks, AWS, Azure, Oracle, Google Cloud, Tableau, Alteryx, or another customer-selected external platform.

Live IRIS Evaluation Smoke Test

This smoke test proves the delivered DataAI JARs can read fictional rows from InterSystems IRIS, run DataAI quality processing and matrix balancing, write customer-selected result tables, and read the persisted results back.

Prerequisites

  • Java 17 and Apache Spark 3.5.0 with spark-submit available.
  • A running evaluation IRIS instance and an isolated namespace.
  • A customer-supplied InterSystems JDBC driver compatible with IRIS and Java.
  • An evaluation identity permitted to read Source and create/write DataAI tables in that isolated namespace.

Spark/Hadoop and the InterSystems JDBC driver are not bundled. The ZIP does include every DataAI JAR needed by the smoke-test application.

Run

  1. Review the evaluation license in the package root.
  2. In Management Portal SQL, select the isolated namespace and execute each statement in setup-evaluation.sql once.
  3. Run the test without putting a password in a command or shell history:
.\run-smoke-test.ps1 `
-IrisJdbcJar 'C:\InterSystems\IRIS\dev\java\lib\1.8\intersystems-jdbc-3.10.1.jar' `
-JdbcUrl 'jdbc:IRIS://localhost:1972/USER'
  1. Enter the evaluation user and password when prompted.
  2. Require the marker DATAAI_IRIS_SMOKE_TEST_PASS and a zero exit code.
  3. Optionally execute verify-results.sql in the same namespace.

The expected fictional-data result is six input rows, four accepted rows, two rejected rows, two quality findings, and four converged matrix cells.

Verify the downloaded ZIP

$expected = (Get-Content .\DataAI_ETL_IRIS_Evaluation.zip.sha256).Split()[0]
$actual = (Get-FileHash .\DataAI_ETL_IRIS_Evaluation.zip -Algorithm SHA256).Hash.ToLowerInvariant()
if ($actual -ne $expected) { throw 'DataAI ETL package checksum mismatch.' }