How does joining across databases work?

Doe pulls results from each database separately, then joins them in its sandbox using whatever shared keys exist: external IDs, email addresses, or mapping tables you define. Match rates are reported so you know how many records linked and which didn't. This is not a single SQL query across engines.

Can I save a dataset request and rerun it later?

Yes. Every request is saved as a reusable template. Rerun against current data anytime, or schedule recurring runs like "refresh this customer list every Monday."

Does this work across different database types?

Yes. Neon and Supabase (Postgres) and PlanetScale (MySQL) simultaneously. Doe handles SQL dialect differences automatically. Cross-database joins happen in the sandbox, not as a single query.

Can non-technical team members request datasets?

Yes. They describe what they need in plain language. Doe translates it to SQL, executes across your databases, and delivers a spreadsheet. The data team can review the generated queries if needed.

Describe the data. Get the spreadsheet.

Tell Doe "all enterprise customers who signed up this quarter with their usage metrics and contract values." It writes the SQL, pulls from your databases, joins the results in its sandbox, and delivers a clean spreadsheet. Every query is visible. Match rates are reported.

Automate Custom Dataset Builder or talk to sales

Doe lets you describe a dataset in plain language and handles the rest: writes the SQL, pulls from your Postgres and MySQL databases, joins the results in a sandbox, and delivers a clean Google Sheets spreadsheet. Connects to Neon, Supabase, and PlanetScale. Every query is visible, match rates are reported for each join, and requests are saved as reusable templates.

What changes

Dimension	Before	With Doe
Request to delivery	Hours to days depending on analyst availability	Minutes, depending on dataset complexity
Repeat requests	Start from scratch ("can you rerun that with Q1 data?")	Saved as a template, rerun anytime with current data
Cross-source joins	CSV exports and VLOOKUP, losing rows from ID mismatches	Doe pulls from each database and joins in its sandbox with match rates reported
Traceability	A spreadsheet with numbers and no record of how they were derived	Metadata tab with every query, match rate, and execution log

How Doe builds custom datasets

Maps the plain-language request to tables and join keys

Doe resolves "enterprise customers with usage and contract data" to specific tables across your databases and generates a query plan you can review before execution.

Pulls product usage metrics via Postgres

Neon

Pulls customer records and contract data via MySQL

PlanetScale

Doe joins results from both databases in its sandbox. Match rate reported so you see how many records linked and which didn't.

Writes the clean spreadsheet with metadata

Google Sheets

Doe delivers named columns, summary statistics, and a metadata tab documenting every query, execution time, and row counts at each join stage.

Every stakeholder needs "just one quick data pull." You have twelve in the queue.

Sales needs enterprise customers with usage data by 3 PM. Marketing wants churned accounts with their last campaign touch. The CEO needs revenue by cohort and geography for a board slide. Each request is quick in isolation. Together, they're your entire week. And every one requires joining data that lives in different databases with different ID formats and a mapping table nobody documented.

The urgent ones get done with CSV exports and VLOOKUP, losing rows where IDs don't match across systems. The non-urgent ones sit in a Jira backlog growing faster than you can clear it. The analysis you were actually hired for gets pushed to "next sprint" indefinitely.

Get started in under 10 minutes

Connect your tools

One-click OAuth for each integration. No API keys, no engineering.

Describe what you need

“Give me all enterprise accounts that signed up in Q4 with their monthly usage, contract value, and support ticket count. Export it as a Google Sheet with one row per account.”

It runs on schedule

On demand, or set a schedule so the dataset refreshes itself and lands in your spreadsheet.

Custom Dataset Builder FAQ

Yes. Doe generates a query plan mapping your request to specific tables and columns. You can review the SQL, edit it, or approve as-is before execution. Every query and its results are logged.

Related workflows

Data & Analytics

Automated KPI Report