Summary
A project does not own databases: it links the workspace’s databases it needs. The project’s Data Warehouse › Databases page lists those linked databases; each opens on its detail page, read-only. Unlinking a database deletes nothing.
The databases linked to the project
A database is declared once, in the workspace’s warehouse — that is where it is created, edited, exported and versioned, see The workspace’s databases. The project links to it. Several projects can therefore work on the same database without duplicating it.
The project’s page shows the same cards as the warehouse: the database’s name, its status, its engine and schema, and the patient count once known. A search and a status filter help once the list grows.
Databases
Databases linked to this project provide the data for its warehouse and analyses.
MIMIC-IV demo (OMOP)
Connected · Duckdb / OMOP CDM 5.4
100 patients
Linking a database
The Link database button opens the Link a database dialog. It only offers the databases of the same workspace that are not linked to the project yet; clicking one links it.
If there are none, the dialog says so — “No databases available to link”. The New database button at the bottom of the dialog then opens the Add a database dialog: the database is created in the workspace and linked to the project straight away. Its configuration is described in Adding a database.
Unlinking a database
A card’s … menu has a single entry: Unlink. Once confirmed, the database leaves the project but stays in the warehouse, untouched, for the other projects. To unlink several, select the cards with ⌘/Ctrl + click or Shift + click, then Unlink.
Linking and unlinking require the editor role on the project.
Deleting a database happens in the workspace
A project offers neither Edit, Export nor Delete: these act on the database itself, shared by every project that links to it. They are in the workspace’s Data Warehouse › Databases.
Consulting a database from the project
Clicking a card opens the database’s detail page, the same as in the workspace, but read-only and on five tabs: Overview, Statistics, Schema, Mapping and SQL.
Statistics have not been computed yet
The counts below read "—" until you load them. On large databases this can be slow, so they are only computed on request.
Readme
MIMIC-IV demo (OMOP)
A demo extract of MIMIC-IV converted to OMOP CDM 5.4: 100 patients, their ICU stays, their measurements and their prescriptions.
- Source: MIMIC-IV Clinical Database Demo 2.2 (PhysioNet)
- Dates shifted per patient, as in MIMIC
- Vocabularies: SNOMED, LOINC, RxNorm, ICD-10
About
Connection
/srv/linkr/data/databases/mimic_iv_demo.duckdb412.3 MB
What changes compared with the workspace:
- No Cohorts tab — a database’s own cohorts belong to the warehouse. The project’s cohorts are built in Cohorts.
- No More menu — the readme, license, export and versioning describe the database: they stay in the workspace. The Overview shows the readme, without an Edit button.
- A read-only mapping — overrides belong to the database, see A database’s mapping.
The rest reads as in the workspace: statistics are computed on request with Load statistics, the schema browser gives the columns and their statistics, and the SQL console queries the database. It is all described in The detail page.
Choosing a page’s database
A project can link several databases — a raw database and its OMOP version, two hospitals. The project’s pages that read data then ask which one: Concepts has a database picker in its toolbar, a cohort is created on a database picked in its creation dialog, and so is a Patient data board. These pickers only offer the databases linked to the project.
Going further
- The workspace’s databases — creating a database, attaching a schema, reading its detail page.
- Concepts — what a linked database actually holds.
- Cohorts — selecting patients in a linked database.