Guides an AI agent through adding data sources to Dremio by asking the user the right questions, recommending connection settings, and linking to the exact documentation for each connector.

2026-03-060

dremio-iceberg-operations

developer-advocacy-dremio/dremio-skills-bundle

Teaches an AI agent how to create, manage, and maintain Apache Iceberg tables in Dremio — including DML, schema evolution, time travel, table maintenance, partitioning, and versioned catalog workflows.

2026-03-060

dremio-lakehouse-modeling

developer-advocacy-dremio/dremio-skills-bundle

Teaches an AI agent data modeling best practices in a Dremio lakehouse — medallion architecture, views vs tables, reflections strategy, partitioning, dimensional modeling, and semantic layer design.

2026-03-060

dremio-python-libraries

developer-advocacy-dremio/dremio-skills-bundle

Enables an AI agent to use dremio-simple-query (lightweight SQL/Arrow Flight) and DremioFrame (full dataframe builder with ingestion, modeling, and admin) to interact with Dremio from Python.

2026-03-060

dremio-sql-reference

developer-advocacy-dremio/dremio-skills-bundle

Teaches an AI agent the Dremio SQL dialect — unique syntax, Iceberg DML, reflections DDL, versioned queries, RBAC commands, and links to the full SQL reference docs.

2026-03-060

name	Dremio CLI
description	Enables an AI agent to install, configure, and use the dremio-cli Python tool to manage Dremio Software and Cloud from the command line.

Dremio CLI Skill

This skill equips you to use the dremio-cli — a Python CLI tool that provides 100% API coverage for both Dremio Software (self-hosted) and Dremio Cloud. Use it to manage catalogs, execute SQL, manage sources/views/jobs, control access, and perform GitOps workflows via Dremio-as-Code.

GitHub: https://github.com/developer-advocacy-dremio/dremio-python-cli
Full Docs: https://github.com/developer-advocacy-dremio/dremio-python-cli/blob/main/dremio-cli/docs/README.md

Installation

Requires Python 3.8+.

# Standard install
pip install dremio-cli

# Or use pipx for an isolated environment
pipx install dremio-cli

# Verify
dremio --version

Credential Configuration

The CLI supports profile-based configuration and environment variables. Ask the user which Dremio edition they use (Software or Cloud) and guide them through setup.

Option 1: Profiles (Recommended)

Profiles are stored locally and can be switched between. A user can have multiple profiles for different environments.

Dremio Software Profile

dremio profile create --name myprofile --type software \
  --base-url https://dremio.company.com \
  --username admin --password secret

Dremio Cloud Profile

dremio profile create --name cloud-prod --type cloud \
  --base-url https://api.dremio.cloud \
  --project-id <project-id> \
  --token <pat-token>

For the EU control plane, use --base-url https://api.eu.dremio.cloud.

Switching Profiles

# Use a specific profile for a single command
dremio --profile cloud-prod catalog list

# Set a default profile
dremio profile set-default myprofile

Option 2: Environment Variables

The user can also set credentials via environment variables or a .env file:

export DREMIO_BASE_URL=https://dremio.company.com
export DREMIO_USERNAME=admin
export DREMIO_PASSWORD=secret

Agent prompt: "To use the Dremio CLI, I'll need to configure credentials. Are you using Dremio Software or Dremio Cloud? Would you like to create a named profile or set environment variables?"

Output Formats

The CLI supports three output formats. Use --output json when you need to parse results programmatically.

# Table (default, human-readable)
dremio catalog list

# JSON (best for scripting and parsing)
dremio --output json catalog list

# YAML
dremio --output yaml catalog list

Use --verbose for debugging.

Command Reference

Catalog Operations

dremio catalog list                    # List all catalog items
dremio catalog get <id>                # Get item details by ID
dremio catalog get-by-path <path>      # Get item details by path

SQL Operations

dremio sql execute "<query>"           # Execute a SQL query
dremio sql explain "<query>"           # Show execution plan
dremio sql validate "<query>"          # Validate SQL syntax

SQL execution is asynchronous on Dremio Software — the CLI handles polling automatically and returns results.

Use --async for long-running queries to return the job ID immediately without waiting.

Source Management

dremio source list                     # List all sources
dremio source create --name MyDB \
  --type POSTGRES --config-file db.json  # Create a source from config
dremio source refresh <id>             # Refresh source metadata

View Management

dremio view list                       # List views
dremio view create --path "Analytics.sales_summary" \
  --sql "SELECT date, SUM(amount) FROM sales GROUP BY date"
dremio view update <id>                # Update a view

Job Management

dremio job list                        # List jobs
dremio job list --max-results 10       # List recent jobs (limited)
dremio job get <job-id>                # Get job details
dremio job results <job-id>            # Get job results
dremio job cancel <job-id>             # Cancel a running job
dremio job profile <job-id> \
  --download profile.zip               # Download job profile

Space & Folder Management

dremio space create --name Analytics   # Create a space
dremio folder create --path <path>     # Create a folder

Access Control (Grants, Users, Roles)

dremio grant list <id>                 # List grants on an object
dremio grant add <id> \
  --grantee-type ROLE \
  --grantee-id analyst \
  --privileges SELECT                  # Grant access
dremio user list                       # List users
dremio role list                       # List roles

Tags & Wiki

dremio wiki set <id> --file README.md  # Set wiki content from file
dremio tag set <id> --tags "production,sensitive,pii"  # Set tags

Platform Support Matrix

Feature	Software	Cloud
Catalog Operations	✅	✅
SQL Execution	✅	⚠️ Limited
Job Management	✅	✅
View Management	✅	✅
Source Management	✅	✅
Space/Folder Mgmt	✅	✅
Tags & Wiki	✅	✅
Grant Management	✅	✅
User Management	✅	⚠️ Via Console
Role Management	✅	⚠️ Via Console
Table Operations	✅	✅

Dremio-as-Code (GitOps)

The CLI includes a sync feature for managing Dremio catalog objects (Spaces, Folders, Views) as local files — enabling GitOps workflows.

Setup

Create a dremio.yaml in your project root:

version: "1.0"
scope:
  path: "dremio-catalog.finance"   # Dremio path to sync
  type: "SPACE"                    # SPACE or ICEBERGCATALOG
ignore:
  - "*.tmp"

Pull/Push Workflow

# Pull current state from Dremio into local files
dremio sync pull

# Edit views locally (SQL + YAML pairs)
# Then push changes back to Dremio
dremio sync push

Pulling creates a directory structure mirroring Dremio with .sql and .yaml file pairs for each view.

View Definition Format

Each view is defined by a YAML file that can include SQL, tags, wiki, access control, governance policies, and reflections:

name: revenue_report
type: VIRTUAL_DATASET
path: ["dremio-catalog", "finance", "reports", "revenue_report"]
sql: |
  SELECT region, sum(amount) as total 
  FROM "dremio-catalog".finance.stg_sales 
  GROUP BY region
dependencies:
  - "stg_sales"
tags: ["finance", "official"]
description: "docs/revenue_report.md"

access_control:
  roles:
    - name: "finance_managers"
      privileges: ["SELECT"]
  users:
    - name: "auditor@example.com"
      privileges: ["SELECT", "ALTER"]

Note: Governance policies (RBAC, row access, masking) and reflections are NOT automatically pulled — they must be manually defined in YAML. Full DAC docs: https://github.com/developer-advocacy-dremio/dremio-python-cli/blob/main/dremio-cli/docs/dac.md

Common Workflows

Data Pipeline Setup

# 1. Create a source
dremio source create --name MyDB --type POSTGRES --config-file db.json

# 2. Create a space
dremio space create --name Analytics

# 3. Create a view
dremio view create --path "Analytics.sales_summary" \
  --sql "SELECT date, SUM(amount) FROM sales GROUP BY date"

# 4. Grant access
dremio grant add <view-id> --grantee-type ROLE \
  --grantee-id analyst --privileges SELECT

Monitoring Jobs

dremio job list --max-results 10
dremio job get <job-id>
dremio job profile <job-id> --download profile.zip

Documenting Datasets

dremio wiki set <id> --file README.md
dremio tag set <id> --tags "production,sensitive,pii"

Agent Workflow

When the user asks you to manage Dremio using the CLI:

Check installation — Run dremio --version. If not installed, install with pip install dremio-cli.
Check configuration — Run dremio profile list to see existing profiles. If none exist, guide the user to create one (Software vs Cloud).
Execute commands — Use the command reference above. Prefer --output json when you need to parse results.
Look up advanced docs — If you need details on a specific command group, read the relevant doc page from the repo:

Tips

Use --output json when you need to extract IDs or specific fields from command output.
Use --verbose to debug connectivity or authentication issues.
Use --async for SQL execution when you only need to submit a query and get the job ID back.
Store SQL queries in files and reference them rather than inlining long queries.
Profiles are stored in the user's home directory — multiple profiles can coexist for different environments (dev, staging, prod).