Introducing Data Connections!

[softwarenerd]

Introducing Data Connections (Experimental Preview)

We're excited to share an early look at Data Connections, a new experimental feature in Positron, and we'd love your feedback.

Data Connections brings the existing Connections and Catalog Explorer features together into a single, unified experience. Our goal is for Data Connections to become the way you work with data in Positron, from local files and database servers to cloud data warehouses and the data catalogs that govern them.

For now, the existing features remain unchanged. Data Connections, Connections, and Catalog Explorer all exist side-by-side, and nothing changes about how Connections and Catalog Explorer work today. In a future release, once Data Connections is more complete, it will replace both Connections and Catalog Explorer.

This is an early, experimental preview. It's turned off by default, the set of supported databases is still small, and the experience will change as we build it out. We're sharing it now because we want you to try it and tell us what works, what doesn't, and what you need.

What's in this first preview

Data Connections currently supports three database drivers:

• DuckDB -- connect to DuckDB database files
• PostgreSQL -- connect to PostgreSQL database servers (local and remote)
• SQLite -- connect to SQLite database files

More drivers are on the way. We plan to add support for Databricks, Snowflake, Redshift, Spark, and others over the coming releases.

Once connected, you can:

• Browse your schemas, tables, views, columns, and indexes in the Data Connections tree
• Double-click a table, view, or column to open it in the Data Explorer
• Generate ready-to-run connection code in Python or R to use in your own scripts and notebooks

Turning on Data Connections

Data Connections is turned off by default. To turn it on:

1. Open Settings (Cmd+, on macOS, Ctrl+, on Windows/Linux).
2. Search for Data Connections.
3. Turn on Data Connections: Enabled.

Or add this line to your settings.json file:

"dataConnections.enabled": true

Then reload Positron for the change to take effect (Command Palette -> Developer: Reload Window).

This is an application-level setting, so it applies across all your Positron windows and appears under the User tab in Settings (not Workspace).

After reloading, you'll find a new Data Connections view in the side bar.

Creating a connection

Open the Data Connections view and click Add Connection (+ icon). From there:

1. Choose a provider -- pick DuckDB, PostgreSQL, or SQLite.

2. Choose how to connect (PostgreSQL only) -- PostgreSQL offers several connection methods; the other drivers skip this step.

3. Fill in the details and give your connection a name.

4. Click Save. Your connection appears in the tree, ready to browse.

Below are the specifics for each database.

DuckDB

1. Click Add Connection and select DuckDB.
2. In Database File, choose the .duckdb file you want to open. (Paths can be absolute, relative to your workspace, or start with ~/.)
3. Optionally turn on Read Only.
4. Name the connection and click Save.

SQLite

1. Click Add Connection and select SQLite.
2. In Database File, choose your .sqlite / .db file. (Paths can be absolute, relative to your workspace, or start with ~/.)
3. Optionally turn on Read Only.
4. Name the connection and click Save.

The Read Only option (DuckDB and SQLite) opens the database without write access. It has little effect today, since connections are read-only for browsing -- it's groundwork for a future release where you'll be able to run your own SQL against a connection.

PostgreSQL

PostgreSQL supports several ways to connect. After selecting PostgreSQL, choose the method that matches your setup:

• User & Password -- connect to a server over the network with a host, port, database, user, and password. Optionally enable SSL. (If you leave the password blank, Positron can use your .pgpass file or the PGPASSWORD environment variable.)
• Local Server (No Password) -- (macOS and Linux only) connect to a PostgreSQL server running on your own machine using your operating system account. Just provide the database name. On Windows, use User & Password to connect to a local server.
• Client Certificate (SSL) -- connect over SSL and authenticate with a client certificate (CA certificate, client certificate, and client key).
• Connection String -- paste a connection string (URL) from your database provider, for example postgresql://user:password@host:5432/database.

Fill in the fields for your chosen method, name the connection, and click Save.

Passwords, connection strings, and other secrets are stored securely in your operating system's secret storage, not in plain-text settings.

Managing connections

Your saved connections persist across sessions and appear in the Data Connections tree. Use a connection's ··· button on the right (or right-click it) to manage it:

• Edit Connection -- reopen the configuration to change its settings.
• Remove -- delete the saved connection.

A Test Connection option is coming in the next release, so you can verify your settings while configuring a connection.

Generating connection code

Open a connection's ··· menu (or right-click it) and choose Connect With -> Python or Connect With -> R to generate connection code you can drop into your own work. Positron generates idiomatic snippets:

• Python -- using native drivers (psycopg2, sqlite3, duckdb) or SQLAlchemy
• R -- using DBI with the appropriate backend (RPostgres, RSQLite, duckdb)

For example:

The generated code imports packages that must be installed in your active Python or R environment -- for example, psycopg2 or sqlalchemy for Python, or DBI and RPostgres for R. Install the package for your database and language before running the snippet.

What's not here yet

This is an early preview, so some things are still on the way:

• Running your own SQL. Browsing is read-only today; executing arbitrary SQL against a connection is coming in a future release.
• SQL code generation. Connection code is currently Python and R only.
• Automatic dependency setup. Today you install the required Python or R packages yourself; we plan to detect and install them for you in a future release.
• Clearer error messages. When a connection fails, today's errors can be terse or come straight from the underlying driver; we're working on clearer, more actionable messages.

We want your feedback

This is just the beginning, and your input will directly shape where it goes. We'd especially love to hear:

• Which databases, data warehouses, and connection methods you most need us to support next
• Whether the connection setup flow is clear and complete for your environment
• Anything that's confusing, missing, or broken

Please reply to this discussion with your thoughts, or file an issue if you hit a bug. If you would like to talk to our product team about Data Connections and SQL support in Positron, schedule a call here.

Thanks for trying it out!

[jonpeake]

This is great! Any plans for ROracle support on the horizon?

[cindyytong]

Hello @jonpeake I've logged your request in this issue #14683. If you're open to chatting with us live, we'd like to learn more about your current workflow managing connections and working with SQL.

[cpilat97]

Would like to echo Jon's comment here. I use oracle quite a bit and have the VS Code Oracle SQL Developer extension side loaded but would be great to use with the native data connections explorer

[pbosetti]

Great stuff. I work with MongoDB (authored the mdbplyr CRAN package) and I'd like to see MongoDB in the list of supported connection. I know it's NoSQL, but Spark is NoSQL too.

[juliasilge]

Thank so much! I've logged this request in #14799. If you are interested in sharing more about how you'd plan to work with MongoDB in Positron and what features would be helpful, we'd love to chat with you live.

[henningsway]

Looks very interesting!
To me Databricks and Postgres Connections are most commonly used currently and therefore most valuable.
Please keep up the good work. :)

[henningsway]

PS: I just realised, that there is no "disconnect" button, e.g. to shut down a remote sql warehouse to reduce costs. From my first impressions the (databricks) connection seems to be upheld beyond the timeout of the warehouse even if it's not used in positron.

But the connections itself seem to work fine! :)

[juliasilge]

Thank you! I re-opened #14623 for this, because we still didn't actually get a way to disconnect from the linked PR.

[henningsway]

Just another kind of thought: I feel the Data Connections should be easily integrated with the Posit Assistant so that the Assistant is aware of which connections are there, what they contain and also be able to initiate a session (without me having to start one interactively). I'm sure you have sth like this in mind already, though!

[henningsway]

• one more: I do feel that this is a nice interface to explore pins as well. I don't have access to connect right now, but I do like to use the package with a different backend (e.g. data bricks volumes). Ideally, I could see/explore all the (versioned) pins, see their metadata, easily load them into the data viewer or my interactive (agent) session.

Tables in some Schema will also work ofc.

[juliasilge]

Thank you for this feedback! 🙌

• We are tracking awareness for Assistant of Data Connections in Data Connections: Add positronDataConnections.connect and executeQuery commands #14924, if you'd like to follow along.
• We are sort of tracking non-Connect pins support in Connections: support other types of blob storage, perhaps as pins #15176, although more as how to help folks deal with blob storage like Databricks Volumes more directly than probably as a pin per se.

[jrosell]

I really miss Mysql/MariaDB here.

[cindyytong]

Thank you for the feedback. We have logged this request here #14855.

[mns-nordicals]

This is very interesting.

I did a quick test with a duckdb file and it seems to have worked well. On the "old" connections pane it is pretty clear when you are connected and when you are disconnected. I can't seem to find a "disconnect" button. I think this is useful information for various reason.

When I collapse the connection (arrow horizontal) and then un-collapse (click) the connection I see a brief spinner. So I guess when the connection is collapsed it gets closed? Perhaps the database icon could be green when it is connected.

[juliasilge]

We've got this logged at #14623 and plan to complete it by the next release!

[ajjitn]

This is niche but since you all already support DuckDB, a Motherduck connection would be awesome!

[juliasilge]

Thanks for this suggestion! I opened #15100 for us to track this as a feature request.

[mns-nordicals]

It seems like the connections you add are global in the sense that when you go into a new project they still show up in that new project. I have a few project where I use duckdb and when I am in a different "project context" I would prefer to only see the connections that are relevant to that project.

Also, I notice that the duckdb connection, where I have specified the path as a relative path - when I switch to a different project it still shows up and doesn't work, which makes sense because now the path is wrong.

So I think you should consider, if it is possible to have some pr project settings and weather database files (duckdb, sqlite) should only accept full paths.

[testlabauto]

Thanks for the suggestion! I have opened #15353 for us to track this as a feature request.

[jeffkeller-einc]

This is a fantastic addition — I've wanted something like this for a long time. First-class SQL/database support feels like one of the last remaining "big" features for Positron.

One thing I noticed after playing with it for a bit: it could use more prominent progress indication. I opened a very slow view in the Data Explorer and saw nothing but a blank window for a few minutes. Eventually the summary sidebar appeared (column names only, no summary statistics), along with row/column counts and a small spinner in the lower-left corner signaling that something was happening.

It would help if that indicator were both more visible and more informative — perhaps naming the current stage (introspecting, running query, fetching data) so I don't have to wonder whether the process has hung.

[juliasilge]

Thank you for this feedback! I opened #15463 for us to track this, and we've definitely heard this from some other venues as well.

[adrian-pascual]

BigQuery would be fantastic as well, thanks!

[juliasilge]

Ah, interesting! 👀 How do you like to use BigQuery now? With what tool? Maybe the bigquery R package?

[maraxe01-molndal]

Thanks for building Data Connections. I’d like to add a request for Microsoft SQL Server support.

A lot of us work in environments where SQL Server is one of the main data sources. Since Positron already supports SQL Server in the existing Connections Pane via pyodbc/pymssql, and R users can connect through DBI + odbc, it would be great to have it in the new Data Connections experience too.

For a first version, basic ODBC support, schema/table browsing, and R/Python code generation would already be very useful.

Thanks!

[juliasilge]

Thank you! I added your info to #14854, which is where we're tracking support for odbc drivers.

[jake-wittman]

I'll second this!