Data Imports

Overview

The imports command group manages data import jobs that bring data from external sources into Arize. There are two sub-groups:

imports files — Import from cloud object storage (S3, GCS, Azure Blob Storage)
imports tables — Import from data warehouses (BigQuery, Snowflake, Databricks)

File Import Commands

Command	Description	Client Method
`imports files list`	List file import jobs	`get_all_file_import_jobs`
`imports files get`	Get a file import job by ID	`get_file_import_job`
`imports files create`	Create a file import job	`create_file_import_job`
`imports files delete`	Delete a file import job	`delete_file_import_job`

Table Import Commands

Command	Description	Client Method
`imports tables list`	List table import jobs	`get_all_table_import_jobs`
`imports tables get`	Get a table import job by ID	`get_table_import_job`
`imports tables create`	Create a table import job	`create_table_import_job`
`imports tables delete`	Delete a table import job	`delete_table_import_job`

File Imports

`imports files list`

arize_toolkit imports files list

Lists all file import jobs in the current space.

Example

$ arize_toolkit imports files list
                    File Import Jobs
┌──────────┬──────────┬────────────┬────────────┐
│ id       │ jobId    │ jobStatus  │ createdAt  │
├──────────┼──────────┼────────────┼────────────┤
│ f1       │ j-abc    │ completed  │ 2025-01-10 │
│ f2       │ j-def    │ running    │ 2025-02-01 │
└──────────┴──────────┴────────────┴────────────┘

`imports files get`

arize_toolkit imports files get JOB_ID

Retrieves details for a file import job, including file counts and schema.

Arguments

JOB_ID — The import job ID.

Example

arize_toolkit --json imports files get "j-abc123"

`imports files create`

arize_toolkit imports files create --blob-store STORE --bucket BUCKET --prefix PREFIX --model MODEL --model-type TYPE --schema JSON [OPTIONS]

Creates a new file import job from cloud object storage.

Required Options

--blob-store — Cloud provider: s3, gcs, or azure.
--bucket — Bucket or container name.
--prefix — Object prefix/path within the bucket.
--model — Target model name.
--model-type — Model type: classification, regression, ranking, object_detection, multi-class, or generative.
--schema — Model schema as a JSON string or @filepath.

Optional Options

--model-version — Model version string.
--environment — Environment: production, validation, training, tracing. Defaults to production.
--dry-run — Validate without importing.
--batch-id — Batch ID for validation data.
--azure-tenant-id — Azure tenant ID (Azure only).
--azure-storage-account — Azure storage account name (Azure only).

Schema Format

The schema defines column mappings. Pass as inline JSON or from a file:

{
  "predictionId": "id_column",
  "timestamp": "ts_column",
  "predictionLabel": "pred_column",
  "actualLabel": "actual_column",
  "features": "feature_"
}

Example

arize_toolkit imports files create \
    --blob-store s3 \
    --bucket "my-data-bucket" \
    --prefix "models/fraud/2025-01/" \
    --model "fraud-detection-v3" \
    --model-type classification \
    --schema @schema.json \
    --environment production

`imports files delete`

arize_toolkit imports files delete JOB_ID [--yes]

Deletes a file import job. Prompts for confirmation unless --yes is passed.

Arguments

JOB_ID — The import job ID.

Options

--yes — Skip confirmation.

Example

arize_toolkit imports files delete "j-abc123" --yes

Table Imports

`imports tables list`

arize_toolkit imports tables list

Lists all table import jobs in the current space.

Example

arize_toolkit imports tables list

`imports tables get`

arize_toolkit imports tables get JOB_ID

Retrieves details for a table import job.

Arguments

JOB_ID — The import job ID.

Example

arize_toolkit --json imports tables get "j-xyz789"

`imports tables create`

arize_toolkit imports tables create --table-store STORE --model MODEL --model-type TYPE --schema JSON --table-config JSON [OPTIONS]

Creates a new table import job from a data warehouse.

Required Options

--table-store — Provider: BigQuery, Snowflake, or Databricks.
--model — Target model name.
--model-type — Model type: classification, regression, ranking, object_detection, multi-class, or generative.
--schema — Model schema as a JSON string or @filepath.
--table-config — Table configuration as a JSON string or @filepath.

Optional Options

--model-version — Model version string.
--environment — Environment: production, validation, training, tracing. Defaults to production.
--dry-run — Validate without importing.
--batch-id — Batch ID for validation data.

Table Configuration Formats

BigQuery

{
  "projectId": "my-gcp-project",
  "dataset": "ml_data",
  "tableName": "predictions"
}

Snowflake

{
  "accountID": "abc123",
  "schema": "PUBLIC",
  "database": "ML_DB",
  "tableName": "PREDICTIONS"
}

Databricks

{
  "hostName": "my-workspace.databricks.com",
  "endpoint": "/sql/1.0/warehouses/abc",
  "port": "443",
  "catalog": "main",
  "databricksSchema": "default",
  "tableName": "predictions"
}

Example

arize_toolkit imports tables create \
    --table-store BigQuery \
    --model "fraud-detection-v3" \
    --model-type classification \
    --schema @schema.json \
    --table-config '{"projectId":"my-project","dataset":"ml","tableName":"preds"}'

`imports tables delete`

arize_toolkit imports tables delete JOB_ID [--yes]

Deletes a table import job. Prompts for confirmation unless --yes is passed.

Arguments

JOB_ID — The import job ID.

Options

--yes — Skip confirmation.

Example

arize_toolkit imports tables delete "j-xyz789" --yes