Getting Started
Step-by-step tutorial to setup PostgreSQL, Fhirbase and perform first data import.
Video Version of This Tutorial
Installing PostgreSQL
To start using Fhirbase you have to install PostgreSQL database first. Any version above 9.6 should be fine. Please proceed to the section describing operating system you're running.
Docker (cross-platform)
1
$ docker run --name fhirbase-postgres -p=5432:5432 -e POSTGRES_PASSWORD=postgres -d postgres:latest
Copied!
Windows
For Windows the most simple way is to use EnterpriseDB PostgreSQL Installer. Also there is a YouTube video demonstrating the whole installation process.
Linux
Most likely that PostgreSQL will be available through your distribution's package manager. On Ubuntu or Debian, it's as easy as
1
$ sudo apt-get install postgresql postgresql-client
Copied!
For other Linux distributions, please use search functionality of your package manager or just Google the right command.
MacOS (Homebrew)
1
$ brew install postgresql
Copied!
Also you'll need to create default user with some password (we advice to use "postgres" as password for simplicity):
1
$ createuser -s -P postgres
2
Enter password for new role: postgres
3
Enter it again: postgres
Copied!
MacOS (Postgres.app)
As an alternative, there is a Postgres.app project which provides PostgreSQL as a regular MacOS application with common drag-and-drop installation. Please follow instructions at Postgres.app Installation Page and especially make sure you've updated you PATH variable to be able to use
psql command-line tool.When you finished installing it, open Terminal and execute following command to set password for
postgres user:1
$ psql -h localhost -U postgres -c '\password postgres'
2
Enter new password: postgres
3
Enter it again: postgres
Copied!
Checking Postgres connection
After you finished installing Postgres, please check it's running and accepting TCP/IP connections.
You can do this with
psql command-line client or with pgAdmin GUI application. For psql, use following command:1
$ psql -h localhost -p 5432 -U postgres -W postgres
2
Password for user postgres: xxxxxxxxxx
3
psql (10.4, server 9.6.3)
4
Type "help" for help.
5
​
6
postgres=# _
Copied!
If you got
postgres=# prompt waiting for your input, your installation is fine. You can quit psql typing \q followed by Enter.If you got an error like this instead:
1
psql: could not connect to server: Connection refused
2
Is the server running on host "localhost" (127.0.0.1) and accepting
3
TCP/IP connections on port 5432?
Copied!
then it looks like PostgreSQL is either not running or not accepting TCP/IP connections. Please check presence of
postgres process using your operating system's Process Viewer (or Task Manager). If postgres process is running, please consider updating pg_hba.conf file as described in this wiki article.TODO: better troubleshooting guide.
Downloading and installing Fhirbase
Go to Fhirbase Download page and download most recent release of Fhirbase. Make sure you've picked right file, it should match your operating system and CPU architecture. Nightly-build release is being updated after every commit to master branch, so it might be unstable.
When you downloaded the Fhirbase binary, make it executable (Windows users don't need this):
1
$ chmod a+x /path/to/downloads/folder/fhirbase-os-arch
Copied!
Then rename it from
fhirbase-os-arch to just fhirbase and put in some directory listed in you $PATH environment variable. To check that it was properly installed, type following command:1
$ fhirbase help
Copied!
It should output Fhirbase command-line help.
Note to Windows users: Fhirbase is a terminal application and it's supposed to be run inside Command Prompt window. If you'll just double-click it, Command Prompt window will popup and disappear immediately. To be able to use Fhirbase, do following:
- Start Menu => Run... => cmd.exe
- type
cd C:\Path\To\Fhirbase\Exeand press Enter (use real path instead of example one) - type
fhirbaseand press Enter (make sure you've renamed downloaded exe file to justfhirbase)
If you don't know what the
$PATH variable is, it's ok, you can skip this step. But you'll have to type full path to the fhirbase executable every time you'll invoke Fhirbase command. So instead of typing1
$ fhirbase help
Copied!
you'll need to type something like this:
1
$ /Users/xxxx/Downloads/fhirbase help
Copied!
So please remember that if your shell says it cannot find
fhirbase command, most likely fhirbase binary wasn't correctly placed inside $PATH.Creating and initializing database
Next step is to create database where we're going to store FHIR data. Again, we can use
psql command-line client or some GUI client like pgAdmin. Connect to Postgres as you previously did:1
$ psql -h localhost -p 5432 -U postgres -W postgres
Copied!
To create database there is a
CREATE DATABASE statement:1
postgres=# CREATE DATABASE fhirbase;
Copied!
Do not forget to put semicolon at the end of statement. Of course, you can change "fhirbase" to any database name you want.
If you got
CREATE DATABASE text as a response to your command, your database was succesfully created. Type \q to quit psql.Now we have database, but before loading any FHIR data into it we need to create schema - some tables and stored procedures. This operation is performed with
fhirbase init command:1
fhirbase --host localhost -p 5432 -d fhirbase -U postgres -W postgres --fhir=3.3.0 init
Copied!
Let's briefly describe meaning of command-line arguments:
--host localhostspecifies host where Postgres is running-p 5432specifies Postgres port-d fhirbasespecifies name of the database to connect to (don't forget to change it if you named database differently)-U postgresname of Postgres user, "postgres" is fine for most cases-W postgresuser's password--fhir=3.3.0FHIR version you're going to use
If you got no error, database was initialized. Most common errors are connection errors (check your Postgres credentials) and SQL errors (you're trying to initialize non-empty database).
Importing FHIR data
We're ready to import data into our newly created database. We can import data from two different sources:
- local files (ndjson files)
Let's try to use local file first. Fhirbase GitHub repository contains sample file, go ahead and download it. When this file is on your local filesystem, invoke
fhirbase load command:1
fhirbase --host localhost -p 5432 -d fhirbase -U postgres -W postgres --fhir=3.3.0 load -m insert /path/to/downloaded/file/bundle.ndjson.gzip
Copied!
Usually it takes from 25 to 45 seconds to import this file. At the end of the input Fhirbase will output statistics about inserted data:
1
Done, inserted 127454 resources in 27 seconds:
2
ImagingStudy 135
3
AllergyIntolerance 272
4
Claim 17937
5
Goal 1100
6
Encounter 15384
7
Procedure 14109
8
Immunization 6548
9
CarePlan 1544
10
Basic 600
11
Organization 768
12
DiagnosticReport 4143
13
MedicationRequest 2553
14
Patient 600
15
Condition 4142
16
Observation 57619
Copied!
There are two different load modes: insert and copy. Using copy mode in some cases can drastically decrease load time. Also you can load data from Bulk Data API endpoints, FHIR bundles and plain JSON files. To get more information on this, take a look at load command reference.
Running Web UI and invoking SQL queries
Fhirbase contains tiny web server to provide basic Web UI for querying database. You can use it instead of desktop Postgres clients for simplicity. Following command will start Fhirbase webserver on http://localhost:3000​
Fhirbase Demo running locally
1
fhirbase --host localhost -p 5432 -d fhirbase -U postgres -W postgres web
Copied!
1
SELECT COUNT(*), resource#>>'{name,0,given,0}'
2
FROM patient
3
GROUP BY resource#>>'{name,0,given,0}'
4
ORDER BY COUNT(*) DESC;
Copied!
As a result you'll get most popular first names for patients stored in your database.
Last modified 3yr ago