5 RStudio, Posit & VS Code

Use RSnowflake and snowflakeR from desktop R tooling

Keywords

snowflake, R, RStudio, Posit, VS Code, workspace notebooks, snowflakeR, RSnowflake, mlops

5.1 Overview

Most R users work in RStudio, Posit Workbench, Positron, or VS Code. Both packages support that workflow: RSnowflake integrates with the Connections Pane and dbplyr; snowflakeR runs the same ML APIs as in Workspace once Python is configured.

Prerequisite: complete Local R Setup (packages, Python env, authentication). This chapter does not repeat those steps except where an IDE needs a specific setting.

You do not need Workspace Notebooks to use Snowflake from R.

5.2 Learning Objectives

Configure Python/reticulate (conda or pip/venv)
Set up VS Code for an RStudio-like experience
Use the Snowflake VS Code extension for SQL and object browsing
Know Posit Workbench Native App and container-hosted RStudio options
Choose between RSnowflake and Posit’s odbc package with the Snowflake ODBC driver

5.3 RStudio & Posit Workbench

5.3.1 Python for snowflakeR

Complete Local R Setup — choose conda, venv, or sfr_install_python_deps().
Set the interpreter — pick one:
- ~/.Renviron: RETICULATE_PYTHON=/path/to/python (restart R)
- Tools → Global Options → Python — select your venv or conda env
- Per session: reticulate::use_virtualenv(...) or use_condaenv(...) before library(snowflakeR)

5.3.2 RSnowflake Connections Pane

With RSnowflake installed, open Connections in RStudio (or Posit Workbench):

New Connection → Snowflake (or dbConnect(Snowflake(), name = "...")).
Browse databases, schemas, tables, and preview data.
Use dbplyr in scripts or R Markdown.

library(DBI)
library(RSnowflake)
library(dplyr)

con <- dbConnect(Snowflake(), name = "default")
orders <- tbl(con, in_schema("ANALYTICS", "ORDERS")) |>
  filter(order_date >= "2024-01-01") |>
  collect()

5.4 RSnowflake vs `odbc` + Snowflake ODBC

On desktop RStudio, Posit Workbench, and Positron, you have two well-supported ways to run DBI and dbplyr against Snowflake. Both are good options; many teams standardize on one for operations and keep the other in mind for edge cases.

Layer	RSnowflake	`odbc` + Snowflake ODBC
R package	RSnowflake	CRAN `odbc`
Under the hood	SQL REST API (+ optional ADBC for bulk I/O)	unixODBC / iODBC → Snowflake ODBC driver
Extra install	None beyond R (optional `adbcsnowflake` for Arrow bulk)	OS driver manager + Snowflake ODBC driver (install guide)
Posit-specific drivers	Not required	Posit Professional Drivers (paid products) or Snowflake’s driver from Snowflake’s site — see Posit: Snowflake
Connections pane	New Connection → Snowflake (RSnowflake)	ODBC DSN / driver-based connection (traditional path)
`connections.toml`	Supported via `dbConnect(Snowflake(), name = "...")`	Supported via `dbConnect(odbc::snowflake(), connection_name = "...")`
dbplyr / DBI	Yes	Yes — same verbs, different driver
snowflakeR (ML APIs)	Orthogonal — install Python + snowflakeR either way	Same

5.4.1 Practical differences to plan for

Topic	RSnowflake	`odbc` + Snowflake ODBC
Admin / CI footprint	R package (+ optional ADBC build for bulk)	Driver binaries, `odbc.ini` / DSN, often IT-managed
Stage file upload/download	The SQL REST API does not support Snowflake `PUT` and `GET` commands (upload/download files to internal stages). Use ADBC bulk table loads, SQL `COPY`, or stage volume mounts in containers instead (Query — limits)	ODBC/JDBC drivers often support stage-oriented file workflows teams already use
Writing large data.frames to tables	RSnowflake `upload_method = "auto"` (default) chooses the fastest path: ADBC bulk upload when installed, else Snowpark in Workspace, else batched SQL INSERT (Query — bulk)	Depends on driver and DSN; no RSnowflake-specific routing
Workspace Notebooks	First-class: session OAuth via `SNOWFLAKE_HOST`; default without ODBC install	ODBC was impractical in notebooks until CRE — driver + unixODBC do not fit `setup_notebook()`; bake into CRE if you must use `odbc`
Workspace bulk I/O	Optional ADBC in CRE (Snowflake’s optimized Arrow path)	ODBC in CRE possible; not the direction Snowflake optimizes for new bulk work
WebR / WASM	HTTPS SQL API works without a native driver	Needs ODBC in the environment — usually impractical
Identifier case	Defaults to uppercase (ODBC-aligned); `RSnowflake.identifier_case`	Snowflake ODBC behaviour; same quoting pitfalls
Ecosystem age	Snowflake Labs, SQL API era	Long-running Posit + enterprise ODBC playbooks

Neither package replaces snowflakeR for Feature Store, Model Registry, or Snowpark ML — you still add Python + snowflakeR when you need those APIs.

5.4.2 Example: `odbc::snowflake()`

The odbc package provides a Snowflake-specific connector that reads the same ~/.snowflake/connections.toml as the Python connector and can pick up ambient OAuth where the platform provides it:

library(DBI)
library(odbc)

con <- dbConnect(odbc::snowflake(), connection_name = "default")

library(dplyr)
tbl(con, in_schema("ANALYTICS", "ORDERS")) |>
  filter(ORDER_DATE >= "2024-01-01") |>
  head(10) |>
  collect()

Interactive SSO on a laptop:

con <- dbConnect(
  odbc::snowflake(),
  account = "xy12345.us-east-1",
  authenticator = "externalbrowser"
)

5.4.3 When teams pick which

Prefer RSnowflake when…	Prefer `odbc` when…
You want no ODBC install on analyst laptops	IT already rolls out the Snowflake ODBC driver to all machines
You align with Snowflake-Labs R + Workspace bootstrap in this guide	You rely on Posit Connect credential patterns (`connectcreds`) or legacy ODBC DSNs
You need SQL API semantics documented in chapter 11	You need driver-level stage/file workflows from R that assume ODBC
WebR or minimal client footprint matters	You have years of validated ODBC runbooks in production

Many teams use RSnowflake for new notebooks and dplyr, and keep ODBC for Shiny on Connect, older projects, or platforms where the driver is pre-installed (e.g. some Posit Workbench images).

5.4.4 Workspace: why SQL API won, and what CRE changes

Historically in Snowflake Workspace, the odbc package plus Snowflake ODBC driver was a poor fit: installing and configuring unixODBC and the driver at notebook runtime was slow, brittle, and hard to secure compared with HTTPS. That is a major reason RSnowflake (SQL API) exists for in-notebook R. ADBC did not share ODBC’s deployment shape — it installs as an R/Go driver package (bootstrap or CRE) and is Snowflake’s preferred Arrow-native path for large bulk I/O alongside the SQL API.

Path	Workspace without CRE	Workspace with CRE
RSnowflake (SQL API)	Bootstrap or tarball	Pre-baked in image — skip bootstrap
ADBC	Bootstrap (`languages.r.addons.adbc`) + EAI for Go modules	Pre-baked — recommended for bulk
`odbc` + ODBC driver	Impractical via bootstrap	Possible if platform team installs driver + DSN in the CRE image — not via `setup_notebook()`

See RSnowflake in Workspace — bootstrap.

5.4.5 snowflakeR in RStudio projects

my-ml-project/
├── .Renviron              # RETICULATE_PYTHON=...
├── R/
│   ├── 00_connect.R
│   ├── 01_features.R
│   └── 02_train.R
└── my-ml-project.Rproj

5.4.6 Positron

Positron uses the same R + Python model. Configure RETICULATE_PYTHON in .Renviron; Connections Pane behavior matches RStudio when enabled.

5.5 Posit Workbench Native App on Snowflake

Posit Workbench is available as a Native App on Snowflake — a full RStudio/Posit IDE inside your Snowflake account, with data locality and in-account auth.

Aspect	Local RStudio	Posit Native App
Location	Your laptop	Snowflake account
Auth	`connections.toml` / SSO	Inherited Snowflake session
Packages	Install RSnowflake + snowflakeR	Same — install in Workbench session
Best for	Day-to-day dev	Teams requiring in-account IDE

snowflakeR and RSnowflake work the same way: install R packages in the Workbench session, configure Python for snowflakeR, connect with standard APIs.

Note

Deploying and licensing Posit Native App is a separate product process — this guide only covers R package usage once Workbench is available.

5.6 RStudio Server on SPCS (custom container)

Run RStudio Server as a custom SPCS service when you need full RStudio UX inside Snowflake but do not use Posit Workbench Native App.

Pros: native R session, Connections pane, custom image
Cons: you operate Docker, compute pool, ingress, and scaling

Full walkthrough: RStudio Server on SPCS — deploy kit in snowflakeR/inst/rstudio-spcs, vignettes rstudio-spcs and spcs-custom-services.

Posit Workbench Native App (above) is the managed Marketplace alternative.

5.7 VS Code

5.7.1 Core R extensions

Extension	Purpose
R	R terminal, run selection/chunk, help
R Debugger	Breakpoints, step debugging (optional)
Quarto	`.qmd` authoring if you use Quarto
R Syntax	Syntax highlighting (often bundled)

Install languageserver in R for linting/hover (optional but improves IDE feel):

install.packages("languageserver")

VS Code R extension settings (example):

{
  "r.rterm.option": ["--no-save", "--no-restore"],
  "r.lsp.enabled": true,
  "python.defaultInterpreterPath": "${env:HOME}/.virtualenvs/r-snowflakeR/bin/python",
  "[r]": {
    "editor.formatOnSave": false
  }
}

Use pip/venv or conda path in python.defaultInterpreterPath — see Local R Setup.

5.7.2 Snowflake extension

Install the Snowflake Extension for VS Code:

Account explorer (databases, schemas, tables)
SQL worksheets against Snowflake
Complements RSnowflake in R scripts — browse in sidebar, analyze in R

Workflow: Snowflake extension for exploration and ad hoc SQL; RSnowflake/dbplyr in .R files for reproducible pipelines; snowflakeR for ML APIs.

5.7.3 Local Jupyter with R (optional)

IRkernel::installspec() for notebook UI locally — distinct from Snowflake Workspace.

5.8 Terminal R

Rscript batch_score.R

Ideal for CI/CD with key-pair auth in TOML.

5.9 Choosing local IDE vs Workspace

Prefer local IDE / Posit App when…	Prefer Workspace when…
Team uses RStudio/Posit daily	Notebook-first, in-account Git
Posit Native App deployed	Mixed Python+R cells with Snowpark
Heavy Connections Pane / Rmd	Strict no-credential-on-laptop policy

Many teams use both: develop in RStudio or Posit Native App, schedule via Tasks/SPCS.

5.10 Same packages, same objects

Feature views, registry models, and tables written from R are visible to R and Python users — regardless of IDE.

5.11 Next steps

Workspaces overview — Snowflake Notebooks.

SQL API → RSnowflake: Connect.

MLOps on Snowflake → snowflakeR: Connect.

--- title: "RStudio, Posit & VS Code" subtitle: "Use RSnowflake and snowflakeR from desktop R tooling" --- ## Overview Most R users work in **RStudio**, **Posit Workbench**, **Positron**, or **VS Code**. Both packages support that workflow: RSnowflake integrates with the **Connections Pane** and **dbplyr**; snowflakeR runs the same ML APIs as in Workspace once Python is configured. **Prerequisite:** complete [Local R Setup](../03_local_r_setup/index.qmd) (packages, Python env, authentication). This chapter does not repeat those steps except where an IDE needs a specific setting. You do **not** need Workspace Notebooks to use Snowflake from R. ## Learning Objectives - Configure Python/reticulate (conda **or** pip/venv) - Set up VS Code for an RStudio-like experience - Use the **Snowflake VS Code extension** for SQL and object browsing - Know **Posit Workbench Native App** and container-hosted RStudio options - Choose between **RSnowflake** and Posit's **`odbc`** package with the Snowflake ODBC driver --- ## RStudio & Posit Workbench {#sec-rstudio} ### Python for snowflakeR 1. Complete [Local R Setup](../03_local_r_setup/index.qmd) — choose conda, venv, or `sfr_install_python_deps()`. 2. Set the interpreter — pick one: - **`~/.Renviron`:** `RETICULATE_PYTHON=/path/to/python` (restart R) - **Tools → Global Options → Python** — select your venv or conda env - **Per session:** `reticulate::use_virtualenv(...)` or `use_condaenv(...)` before `library(snowflakeR)` ### RSnowflake Connections Pane With RSnowflake installed, open **Connections** in RStudio (or Posit Workbench): 1. **New Connection → Snowflake** (or `dbConnect(Snowflake(), name = "...")`). 2. Browse databases, schemas, tables, and preview data. 3. Use **dbplyr** in scripts or R Markdown. ```r library(DBI) library(RSnowflake) library(dplyr) con <- dbConnect(Snowflake(), name = "default") orders <- tbl(con, in_schema("ANALYTICS", "ORDERS")) |> filter(order_date >= "2024-01-01") |> collect() ``` --- ## RSnowflake vs `odbc` + Snowflake ODBC {#sec-odbc} On desktop **RStudio**, **Posit Workbench**, and **Positron**, you have two well-supported ways to run **DBI** and **dbplyr** against Snowflake. Both are good options; many teams standardize on one for operations and keep the other in mind for edge cases. | Layer | **RSnowflake** | **`odbc`** + Snowflake ODBC | |-------|----------------|----------------------------| | R package | [RSnowflake](https://github.com/Snowflake-Labs/RSnowflake) | CRAN [`odbc`](https://odbc.r-dbi.org/) | | Under the hood | [SQL REST API](../11_rsnowflake_sql_api/index.qmd) (+ optional **ADBC** for bulk I/O) | **unixODBC** / iODBC → [Snowflake ODBC driver](https://docs.snowflake.com/en/developer-guide/odbc/odbc) | | Extra install | None beyond R (optional `adbcsnowflake` for Arrow bulk) | OS driver manager + Snowflake ODBC driver ([install guide](https://docs.snowflake.com/en/developer-guide/odbc/odbc)) | | Posit-specific drivers | Not required | [Posit Professional Drivers](https://docs.posit.co/pro-drivers/) (paid products) or Snowflake's driver from Snowflake's site — see [Posit: Snowflake](https://solutions.posit.co/connections/db/databases/snowflake/) | | **Connections** pane | **New Connection → Snowflake** (RSnowflake) | ODBC DSN / driver-based connection (traditional path) | | **`connections.toml`** | Supported via `dbConnect(Snowflake(), name = "...")` | Supported via `dbConnect(odbc::snowflake(), connection_name = "...")` | | **dbplyr** / **DBI** | Yes | Yes — same verbs, different driver | | **snowflakeR** (ML APIs) | Orthogonal — install Python + snowflakeR either way | Same | ### Practical differences to plan for | Topic | RSnowflake | `odbc` + Snowflake ODBC | |-------|------------|-------------------------| | **Admin / CI footprint** | R package (+ optional ADBC build for bulk) | Driver binaries, `odbc.ini` / DSN, often IT-managed | | **Stage file upload/download** | The SQL REST API does not support Snowflake **`PUT`** and **`GET`** commands (upload/download files to internal **stages**). Use ADBC bulk table loads, SQL `COPY`, or stage **volume mounts** in containers instead ([Query — limits](../13_rsnowflake_query/index.qmd#sec-limits)) | ODBC/JDBC drivers often support stage-oriented file workflows teams already use | | **Writing large data.frames to tables** | RSnowflake **`upload_method = "auto"`** (default) chooses the fastest path: **ADBC** bulk upload when installed, else Snowpark in Workspace, else batched SQL INSERT ([Query — bulk](../13_rsnowflake_query/index.qmd#sec-bulk)) | Depends on driver and DSN; no RSnowflake-specific routing | | **Workspace Notebooks** | First-class: session OAuth via `SNOWFLAKE_HOST`; default without ODBC install | **ODBC was impractical** in notebooks until **[CRE](../09_custom_runtime_and_ml_jobs/index.qmd)** — driver + unixODBC do not fit `setup_notebook()`; bake into CRE if you must use `odbc` | | **Workspace bulk I/O** | Optional **ADBC** in CRE (Snowflake's optimized Arrow path) | ODBC in CRE possible; not the direction Snowflake optimizes for new bulk work | | **WebR / WASM** | HTTPS SQL API works without a native driver | Needs ODBC in the environment — usually impractical | | **Identifier case** | Defaults to uppercase (ODBC-aligned); `RSnowflake.identifier_case` | Snowflake ODBC behaviour; same quoting pitfalls | | **Ecosystem age** | Snowflake Labs, SQL API era | Long-running Posit + enterprise ODBC playbooks | Neither package replaces **snowflakeR** for Feature Store, Model Registry, or Snowpark ML — you still add Python + snowflakeR when you need those APIs. ### Example: `odbc::snowflake()` The `odbc` package provides a Snowflake-specific connector that reads the same **`~/.snowflake/connections.toml`** as the Python connector and can pick up ambient OAuth where the platform provides it: ```r library(DBI) library(odbc) con <- dbConnect(odbc::snowflake(), connection_name = "default") library(dplyr) tbl(con, in_schema("ANALYTICS", "ORDERS")) |> filter(ORDER_DATE >= "2024-01-01") |> head(10) |> collect() ``` Interactive SSO on a laptop: ```r con <- dbConnect( odbc::snowflake(), account = "xy12345.us-east-1", authenticator = "externalbrowser" ) ``` ### When teams pick which | Prefer **RSnowflake** when… | Prefer **`odbc`** when… | |-----------------------------|-------------------------| | You want **no ODBC install** on analyst laptops | IT already **rolls out** the Snowflake ODBC driver to all machines | | You align with **Snowflake-Labs** R + Workspace bootstrap in this guide | You rely on **Posit Connect** credential patterns (`connectcreds`) or legacy ODBC DSNs | | You need **SQL API** semantics documented in [chapter 11](../11_rsnowflake_sql_api/index.qmd) | You need **driver-level** stage/file workflows from R that assume ODBC | | **WebR** or minimal client footprint matters | You have years of **validated ODBC** runbooks in production | Many teams use **RSnowflake for new notebooks and dplyr**, and keep **ODBC** for Shiny on Connect, older projects, or platforms where the driver is pre-installed (e.g. some Posit Workbench images). ### Workspace: why SQL API won, and what CRE changes Historically in **Snowflake Workspace**, the **`odbc` package plus Snowflake ODBC driver** was a poor fit: installing and configuring unixODBC and the driver at notebook runtime was slow, brittle, and hard to secure compared with HTTPS. That is a major reason **RSnowflake** (SQL API) exists for in-notebook R. **ADBC** did not share ODBC's deployment shape — it installs as an R/Go driver package (bootstrap or CRE) and is Snowflake's **preferred Arrow-native** path for large bulk I/O alongside the SQL API. | Path | Workspace without CRE | Workspace with CRE | |------|----------------------|-------------------| | **RSnowflake** (SQL API) | Bootstrap or tarball | Pre-baked in image — skip bootstrap | | **ADBC** | Bootstrap (`languages.r.addons.adbc`) + EAI for Go modules | Pre-baked — recommended for bulk | | **`odbc` + ODBC driver** | Impractical via bootstrap | **Possible** if platform team installs driver + DSN in the **CRE image** — not via `setup_notebook()` | See [RSnowflake in Workspace — bootstrap](../14_rsnowflake_workspace/index.qmd#sec-order). --- ### snowflakeR in RStudio projects ``` my-ml-project/ ├── .Renviron # RETICULATE_PYTHON=... ├── R/ │ ├── 00_connect.R │ ├── 01_features.R │ └── 02_train.R └── my-ml-project.Rproj ``` ### Positron [Positron](https://posit.co/products/positron/) uses the same R + Python model. Configure `RETICULATE_PYTHON` in `.Renviron`; Connections Pane behavior matches RStudio when enabled. --- ## Posit Workbench Native App on Snowflake {#sec-posit-native} **Posit Workbench** is available as a **[Native App on Snowflake](https://docs.posit.co/)** — a full RStudio/Posit IDE **inside your Snowflake account**, with data locality and in-account auth. | Aspect | Local RStudio | Posit Native App | |--------|---------------|------------------| | Location | Your laptop | Snowflake account | | Auth | `connections.toml` / SSO | Inherited Snowflake session | | Packages | Install RSnowflake + snowflakeR | Same — install in Workbench session | | Best for | Day-to-day dev | Teams requiring in-account IDE | snowflakeR and RSnowflake work the same way: install R packages in the Workbench session, configure Python for snowflakeR, connect with standard APIs. ::: {.callout-note} Deploying and licensing Posit Native App is a separate product process — this guide only covers R package usage once Workbench is available. ::: --- ## RStudio Server on SPCS (custom container) {#sec-rstudio-spcs} Run **RStudio Server** as a custom **SPCS service** when you need full RStudio UX inside Snowflake but do not use Posit Workbench Native App. - **Pros:** native R session, Connections pane, custom image - **Cons:** you operate Docker, compute pool, ingress, and scaling **Full walkthrough:** [RStudio Server on SPCS](../04b_rstudio_spcs/index.qmd) — deploy kit in [snowflakeR/inst/rstudio-spcs](https://github.com/Snowflake-Labs/snowflakeR/tree/main/inst/rstudio-spcs), vignettes `rstudio-spcs` and `spcs-custom-services`. Posit **Workbench Native App** (above) is the managed Marketplace alternative. --- ## VS Code {#sec-vscode} ### Core R extensions | Extension | Purpose | |-----------|---------| | [**R**](https://marketplace.visualstudio.com/items?itemName=REditorSupport.r) | R terminal, run selection/chunk, help | | [**R Debugger**](https://marketplace.visualstudio.com/items?itemName=RDebugger.r-debugger) | Breakpoints, step debugging (optional) | | [**Quarto**](https://marketplace.visualstudio.com/items?itemName=quarto.quarto) | `.qmd` authoring if you use Quarto | | [**R Syntax**](https://marketplace.visualstudio.com/items?itemName=REditorSupport.r-syntax) | Syntax highlighting (often bundled) | Install **languageserver** in R for linting/hover (optional but improves IDE feel): ```r install.packages("languageserver") ``` VS Code R extension settings (example): ```json { "r.rterm.option": ["--no-save", "--no-restore"], "r.lsp.enabled": true, "python.defaultInterpreterPath": "${env:HOME}/.virtualenvs/r-snowflakeR/bin/python", "[r]": { "editor.formatOnSave": false } } ``` Use **pip/venv** or conda path in `python.defaultInterpreterPath` — see [Local R Setup](../03_local_r_setup/index.qmd#sec-python-venv). ### Snowflake extension Install the [**Snowflake Extension for VS Code**](https://docs.snowflake.com/en/user-guide/vscode-ext): - Account explorer (databases, schemas, tables) - SQL worksheets against Snowflake - Complements RSnowflake in R scripts — browse in sidebar, analyze in R Workflow: Snowflake extension for exploration and ad hoc SQL; RSnowflake/dbplyr in `.R` files for reproducible pipelines; snowflakeR for ML APIs. ### Local Jupyter with R (optional) `IRkernel::installspec()` for notebook UI locally — distinct from Snowflake Workspace. --- ## Terminal R {#sec-terminal} ```bash Rscript batch_score.R ``` Ideal for CI/CD with key-pair auth in TOML. --- ## Choosing local IDE vs Workspace {#sec-choose} | Prefer local IDE / Posit App when… | Prefer Workspace when… | |-----------------------------------|-------------------------| | Team uses RStudio/Posit daily | Notebook-first, in-account Git | | Posit Native App deployed | Mixed Python+R cells with Snowpark | | Heavy Connections Pane / Rmd | Strict no-credential-on-laptop policy | Many teams use **both**: develop in RStudio or Posit Native App, schedule via Tasks/SPCS. --- ## Same packages, same objects {#sec-same} Feature views, registry models, and tables written from R are visible to R and Python users — regardless of IDE. ## Next steps {#sec-next} [Workspaces overview](../05_workspaces_overview/index.qmd) — Snowflake Notebooks. [SQL API](../11_rsnowflake_sql_api/index.qmd) → [RSnowflake: Connect](../12_rsnowflake_connect/index.qmd). [MLOps on Snowflake](../15_mlops_on_snowflake/index.qmd) → [snowflakeR: Connect](../16_snowflaker_connect/index.qmd).

5.1 Overview

5.2 Learning Objectives

5.3 RStudio & Posit Workbench

5.3.1 Python for snowflakeR

5.3.2 RSnowflake Connections Pane

5.4 RSnowflake vs odbc + Snowflake ODBC

5.4.1 Practical differences to plan for

5.4.2 Example: odbc::snowflake()

5.4.3 When teams pick which

5.4.4 Workspace: why SQL API won, and what CRE changes

5.4.5 snowflakeR in RStudio projects

5.4.6 Positron

5.5 Posit Workbench Native App on Snowflake

5.6 RStudio Server on SPCS (custom container)

5.7 VS Code

5.7.1 Core R extensions

5.7.2 Snowflake extension

5.7.3 Local Jupyter with R (optional)

5.8 Terminal R

5.9 Choosing local IDE vs Workspace

5.10 Same packages, same objects

5.11 Next steps

5.4 RSnowflake vs `odbc` + Snowflake ODBC

5.4.2 Example: `odbc::snowflake()`