Last updated: May 2026

Introduction

Q: How do I change the delimiter when loading CSV files from SFTP?

Set the delimiter option under source_options, for example a pipe or a tab, when your files do not use commas. Sling reads the file with that delimiter so columns are parsed correctly.

Q: Can Sling load several CSV files from one SFTP directory at once?

Yes. Use a wildcard path such as /path/to/csv_folder/*.csv as the stream and Sling processes every matching file. The {stream_file_name} variable lets you give each file its own target table.

Q: Does Sling support incremental loading for CSV files from SFTP?

Yes. Set the stream mode to incremental and define an update_key so Sling loads only rows newer than the last run. This avoids reloading data that is already in Snowflake.

Q: How can I handle empty values in CSV source files?

Enable empty_as_null under source_options so Sling converts empty fields into NULL values. This keeps the loaded table consistent instead of mixing blank strings with genuine nulls.

Moving data from SFTP servers to Snowflake has traditionally been a complex process requiring multiple tools and custom scripts. Organizations often struggle with challenges such as:

Setting up and maintaining secure SFTP connections
Handling CSV file formats and data type mappings
Managing incremental loads and data transformations
Ensuring reliable and scalable data transfers
Monitoring and troubleshooting data pipelines

Sling simplifies this entire process by providing a unified, efficient solution for transferring data from SFTP servers to Snowflake. In this comprehensive guide, we’ll walk through how to set up and use Sling for this specific use case, covering both command-line operations and YAML-based configurations.

Let’s dive into how you can streamline your SFTP to Snowflake data pipeline using Sling.

Prerequisites

Before we begin, make sure you have the following:

Sling CLI Installation

You can install Sling using various methods depending on your operating system:

# macOS / Linux
curl -fsSL https://slingdata.io/install.sh | bash

# Windows
irm https://slingdata.io/install.ps1 | iex

# Python
pip install sling

Verify the installation by running:

# Check Sling version
sling --version

SFTP Server Requirements
- Host address and port (default is 22)
- Username and password or SSH private key
- Path to the CSV files you want to transfer
- Proper read permissions on the SFTP server
Snowflake Requirements
- Account identifier (e.g., xy12345.us-east-1)
- Username and password
- Database name and schema
- Role with appropriate privileges
- Warehouse for processing
CSV File Considerations
- Consistent file format and structure
- Well-defined column names and data types
- UTF-8 encoding recommended
- Proper delimiter and quote characters

Setting Up Connections

Before we can transfer data, we need to configure both our SFTP and Snowflake connections in Sling. Let’s set up each connection step by step.

SFTP Connection Setup

There are several ways to configure your SFTP connection in Sling:

Using sling conns set Command

# Basic authentication with password
sling conns set MY_SFTP type=sftp host=sftp.example.com user=myuser password=mypassword port=22

# Using SSH private key
sling conns set MY_SFTP type=sftp host=sftp.example.com user=myuser private_key=/path/to/private_key port=22

Using Environment Variables

# Using JSON/YAML format
export MY_SFTP='{type: sftp, host: sftp.example.com, user: myuser, password: mypassword, port: 22}'

Using Sling Environment YAML

Create or edit ~/.sling/env.yaml:

connections:
  MY_SFTP:
    type: sftp
    host: sftp.example.com
    user: myuser
    port: 22
    password: mypassword
    # Or use private key authentication
    # private_key: /path/to/private_key

Snowflake Connection Setup

Similarly, let’s configure the Snowflake connection:

Using sling conns set Command

# Basic configuration
sling conns set SNOWFLAKE type=snowflake account=xy12345.us-east-1 user=myuser password=mypassword database=mydb role=myrole warehouse=mywarehouse

# Using connection URL
sling conns set SNOWFLAKE url="snowflake://myuser:[email protected]/mydb?warehouse=mywarehouse&role=myrole"

Using Environment Variables

# Using connection URL format
export SNOWFLAKE='snowflake://myuser:[email protected]/mydb?warehouse=mywarehouse&role=myrole'

Using Sling Environment YAML

Add to your ~/.sling/env.yaml:

connections:
  SNOWFLAKE:
    type: snowflake
    account: xy12345.us-east-1
    user: myuser
    password: mypassword
    database: mydb
    schema: public
    role: myrole
    warehouse: mywarehouse

Testing Connections

After setting up your connections, it’s important to verify they work correctly:

# Test SFTP connection
sling conns test MY_SFTP

# Test Snowflake connection
sling conns test SNOWFLAKE

# List available connections
sling conns list

Sling Connections UI

These tests ensure that Sling can successfully connect to both your SFTP server and Snowflake instance before attempting any data transfers.

Data Transfer Methods

Sling provides two main approaches for transferring data from SFTP to Snowflake: using CLI flags for quick operations and using YAML configurations for more complex, repeatable workflows.

Using CLI Flags

The CLI approach is perfect for quick, one-off transfers or when you’re testing your setup:

Basic Transfer

# Transfer a single CSV file
sling run --src-conn MY_SFTP --src-stream '/path/to/data.csv' \
  --tgt-conn SNOWFLAKE --tgt-object 'public.my_table' \
  --mode full-refresh

# Transfer multiple CSV files from a directory
sling run --src-conn MY_SFTP --src-stream '/path/to/csv_folder/*.csv' \
  --tgt-conn SNOWFLAKE --tgt-object 'public.{stream_file_name}' \
  --mode full-refresh

Advanced Transfer with Options

# Transfer with column mapping and transformations
sling run --src-conn MY_SFTP --src-stream '/path/to/data.csv' \
  --tgt-conn SNOWFLAKE --tgt-object 'public.my_table' \
  --mode full-refresh \
  --columns '{ "id": "integer", "name": "string", "created_at": "timestamp" }' \
  --transforms '[remove_diacritics]' \
  --src-options '{ "empty_as_null": true, "datetime_format": "YYYY-MM-DD" }' \
  --tgt-options '{ "column_casing": "snake", "add_new_columns": true }'

Using Replication YAML

For more complex scenarios or production workflows, using a YAML configuration file provides better maintainability and version control:

Basic Replication Configuration

Create a file named sftp_to_snowflake.yaml:

source: MY_SFTP
target: SNOWFLAKE

defaults:
  mode: full-refresh

streams:
  '/path/to/customers.csv':
    object: public.customers
    columns:
      customer_id: integer
      name: string
      email: string(50)
      created_at: timestamp

  '/path/to/orders.csv':
    object: public.orders
    columns:
      order_id: integer
      customer_id: integer
      order_date: timestamp
      total_amount: decimal

Advanced Replication Configuration

Here’s a more sophisticated configuration with multiple streams, transformations, and options:

source: MY_SFTP
target: SNOWFLAKE

defaults:
  mode: full-refresh
  primary_key: [id]
  source_options:
    empty_as_null: true
    datetime_format: "YYYY-MM-DD"
  target_options:
    column_casing: snake
    add_new_columns: true

streams:
  '/path/to/customers/*.csv':
    mode: truncate
    object: public.customers
    columns:
      id: integer
      name: string
      email: string
      created_at: timestamp
    transforms:
      - remove_diacritics
      - trim_space
    source_options:
      delimiter: "|"

  '/path/to/orders/*.csv':
    object: public.orders_{stream_table}
    mode: incremental
    columns:
      order_id: integer
      customer_id: integer
      order_date: timestamp
      total_amount: decimal
    update_key: order_date

To run the replication:

# Run the replication configuration
sling run -r sftp_to_snowflake.yaml

Sling Editor UI

Runtime Variables

Sling supports runtime variables that can be used in your YAML configurations for dynamic paths and table names:

{stream_file_name}: The name of the current file being processed
{stream_table}: The table name derived from the file name
{stream_schema}: The schema name if specified in the path

These variables are particularly useful when dealing with multiple files or when you want to maintain a consistent naming convention between your source files and target tables.

Sling Platform Overview

While the CLI is powerful for local development and testing, the Sling Platform provides a comprehensive web-based interface for managing your data operations at scale. Here are the key components:

Platform Components

Connection Management
- Centralized credential storage
- Connection testing and validation
- Role-based access control
- Support for multiple environments
Visual Editor
- YAML configuration editor with syntax highlighting
- Real-time validation
- Template library
- Version control integration
Job Management
- Scheduling and orchestration
- Dependency management
- Error handling and retries
- Monitoring and alerting
Agents
- Distributed execution
- Secure access to data sources
- Resource management
- Auto-scaling capabilities

Platform Interface

The Sling Platform provides an intuitive interface for managing your data operations:

Sling Execution UI

Key features include:

Real-time job monitoring
Detailed execution logs
Performance metrics
Error tracking and debugging tools

Agent Management

Agents are the workers that execute your data operations:

Sling Agent UI

Benefits of using agents:

Run in your own infrastructure
Secure access to your data sources
Support for both development and production environments
Automatic updates and maintenance

Job History and Monitoring

Track all your data operations with comprehensive history and monitoring:

Sling History UI

Features include:

Historical execution data
Success/failure tracking
Performance trends
Resource utilization metrics

Getting Started

Ready to streamline your SFTP to Snowflake data pipeline with Sling? Here’s how to get started:

Install Sling CLI
- Follow the installation instructions for your operating system
- Verify the installation with sling --version
Set Up Your Environment
- Configure your connections using the environment guide
- Test your connections with sling conns test
Create Your First Replication
- Start with a simple CLI command to test the transfer
- Create a replication YAML file for more complex workflows
- Learn about replication modes and options
Explore Advanced Features
- Check out source options for file handling
- Learn about target options for Snowflake optimization
- Understand runtime variables for dynamic configurations
Consider the Sling Platform
- Sign up at app.slingdata.io
- Follow the platform getting started guide
- Deploy agents in your infrastructure

Additional Resources

Community and Support

Join our Discord community
Report issues on GitHub
Contact support at [email protected]

For other SFTP and Snowflake loading patterns, see these guides:

Loading JSON data from SFTP to Snowflake when your source files are JSON instead of CSV.
Transferring SFTP Parquet files to Snowflake for columnar Parquet sources.
Importing SFTP files into PostgreSQL for the same source with a PostgreSQL target.
Loading local CSV data into Snowflake for CSV files that originate on your own machine.
Migrating SQL Server to Snowflake for a database-to-warehouse pipeline into the same target.

FAQ

Set the delimiter option under source_options, for example a pipe or a tab, when your files do not use commas. Sling reads the file with that delimiter so columns are parsed correctly.

Can Sling load several CSV files from one SFTP directory at once?

Yes. Use a wildcard path such as /path/to/csv_folder/*.csv as the stream and Sling processes every matching file. The {stream_file_name} variable lets you give each file its own target table.

How does Sling determine column data types for CSV files?

Sling samples the file and infers each column type automatically. You can override the result by declaring explicit types under the columns key, which is useful when a numeric-looking column should stay text.

Does Sling support incremental loading for CSV files from SFTP?

Yes. Set the stream mode to incremental and define an update_key so Sling loads only rows newer than the last run. This avoids reloading data that is already in Snowflake.

How can I handle empty values in CSV source files?

Enable empty_as_null under source_options so Sling converts empty fields into NULL values. This keeps the loaded table consistent instead of mixing blank strings with genuine nulls.

What is the difference between full-refresh and truncate mode for CSV loads?

Full-refresh drops and recreates the target table on each run, while truncate keeps the existing table and empties it before loading. Truncate preserves table grants and structure, which can matter in production.

Do I need to define column types in the replication YAML?

No, column definitions are optional because Sling infers types from the data. Declaring them gives you tighter control over the target schema and is recommended for production pipelines.

Start simplifying your data operations today with Sling!

Efficiently Transfer CSV Data from SFTP to Snowflake with Sling

Introduction

Prerequisites

Setting Up Connections

SFTP Connection Setup

Snowflake Connection Setup

Testing Connections

Data Transfer Methods

Using CLI Flags

Using Replication YAML

Runtime Variables

Sling Platform Overview

Platform Components

Platform Interface

Agent Management

Job History and Monitoring

Getting Started

Additional Resources

Community and Support

FAQ

How do I change the delimiter when loading CSV files from SFTP?

Can Sling load several CSV files from one SFTP directory at once?

How does Sling determine column data types for CSV files?

Does Sling support incremental loading for CSV files from SFTP?

How can I handle empty values in CSV source files?

What is the difference between full-refresh and truncate mode for CSV loads?

Do I need to define column types in the replication YAML?

Platform

Links

Efficiently Transfer CSV Data from SFTP to Snowflake with Sling

Introduction

Prerequisites

Setting Up Connections

SFTP Connection Setup

Snowflake Connection Setup

Testing Connections

Data Transfer Methods

Using CLI Flags

Using Replication YAML

Runtime Variables

Sling Platform Overview

Platform Components

Platform Interface

Agent Management

Job History and Monitoring

Getting Started

Additional Resources

Community and Support

Related Guides

FAQ

How do I change the delimiter when loading CSV files from SFTP?

Can Sling load several CSV files from one SFTP directory at once?

How does Sling determine column data types for CSV files?

Does Sling support incremental loading for CSV files from SFTP?

How can I handle empty values in CSV source files?

What is the difference between full-refresh and truncate mode for CSV loads?

Do I need to define column types in the replication YAML?