Databases
Connect PostgreSQL to DataxLand
The PostgreSQL connector is a native DataxLand source for businesses whose sales software, ERP or in-house app runs on Postgres. You choose the tables and the sync mode of each - full, incremental by an update column, or CDC reading the change log so deleted rows are caught too. Data is copied as is into the workspace's own warehouse, so dashboards and Si never query your live database each time someone looks.
Free plan with no time limit · No card required · Sample workspace ready in minutes
What you can do with PostgreSQL
- Profit and operations dashboards from your own software's database without writing reports.
- Refresh every 15 or 30 minutes when every source syncs incrementally or with CDC.
- Join orders in the database with ads, marketplaces and a cost sheet in Google Sheets.
- Let Si explore unfamiliar tables before proposing a mapping, with limited read-only SQL.
What data you get
Each stream lands as its own table in the workspace warehouse, as is; personal-data columns are hashed on load.
| Stream (table) | Contents |
|---|---|
| <schema>.<table> | Each table or view you pick becomes its own warehouse table, original column names kept. |
| primary key, cursor column | DataxLand suggests the primary key from the PK constraint and a cursor from a timestamp column such as updated_at. |
| CDC changes | In CDC mode, INSERT, UPDATE and DELETE come from a logical replication slot; rows deleted at the source are deleted in the warehouse. |
Connection details
| Sign-in method | Connection details (API key, account), encrypted at rest |
|---|---|
| Sync modes | Full (reloaded each run) · Incremental (new or changed rows only) · CDC (reads the change log, including deletes) |
| Developer | DataxLand |
| License | proprietary |
Passwords and keys are encrypted at rest. Personal-data columns are hashed on load.
Metrics computed from this source
Real definitions from the analytics packs; dashboards and Si share them.
| Metric | Definition | Unit |
|---|---|---|
| Net revenue | Line value after line discounts and allocated order vouchers, minus refunds to buyers. Refunds are deducted only here. | Currency |
| GMV (gross merchandise value) | Total quantity times unit price of the order lines in valid orders (before discounts, vouchers and refunds). | Currency |
| Valid orders | Orders that were not cancelled (standard status other than cancelled). | Count |
| Average order value | Line revenue after discounts divided by valid orders.= line_net_total / Valid orders | Currency |
| Units sold | Total quantity of products in the order lines of valid orders. | Count |
| Cancellation rate | Cancelled orders divided by total orders.= cancelled_orders / all_orders | Ratio (%) |
| Return rate | Returned orders divided by valid orders.= returned_orders / Valid orders | Ratio (%) |
| Contribution margin | Net revenue minus COGS, marketplace fees, payment fees, shipping and return costs (before advertising).= Net revenue - Cost of goods sold - Marketplace fees - Payment fees - Seller-paid shipping cost - Return cost | Currency |
| On-time delivery rate | On-time deliveries divided by all outbound shipments in the period (by ship date).= ops_on_time_deliveries / Shipments | Ratio (%) |
Connect PostgreSQL in 4 steps
- 1
Prepare a read-only user
Create a Postgres user that can only read the needed schemas. For CDC you need wal_level = logical and the REPLICATION privilege (or your provider's equivalent role).
- 2
Enter connection details
In the workspace choose Add source → PostgreSQL, enter host, port, database, schema (default public), user and password, click Test connection.
- 3
Pick tables and sync modes
Pick tables, then Full, Incremental (cursor, primary key) or CDC per table; Check CDC requirements lists anything missing.
- 4
Review Si's proposal
Si proposes what each table is in your industry's standard model; you approve and build.
Limits and caveats
- The database must be reachable from the internet; private network addresses are blocked for safety.
- CDC needs a primary key (or REPLICA IDENTITY FULL). The slot retains WAL on your server - the source page shows its size, warns above 1 GB and drops the slot after 7 days without sync, with notice.
- Incremental mode cannot see deleted rows; use CDC for that.
- The password is encrypted at rest; Si runs read-only SQL within query and cost limits, and queries over the limit wait for a technical user's approval.
Frequently asked questions
Does DataxLand write to my database?
It writes no data. In CDC mode DataxLand only creates and uses its own replication slot, and drops it when you remove the source or turn CDC off.
Do you support MySQL or SQL Server?
Yes. MySQL, MariaDB, SQL Server and Azure SQL are native DataxLand sources too, loaded in full or incrementally by an update column (rowversion on SQL Server); CDC is PostgreSQL only for now.
Do Neon, Supabase or RDS work?
Yes if they accept outside connections; for CDC the provider must enable logical replication.
Dashboard templates that use this source
Connect PostgreSQL and get verified dashboards
Free plan with no time limit · No card required · Sample workspace ready in minutes