Load

Via the GUI

While logged in to the bclink web-gui as the 'data' user.

Create a new person table [optional]

Viewing the info:

Upload data

Via the CLI

ssh into the VM

Log in to the VM hosting bclink and switch to the user bcos_srv:

ssh XX.YY.ZZ
sudo -s
su - bcos_srv

Setup a working directory

Create a new working directory

mkdir /usr/lib/bcos/MyWorkingDirectory/

Prepare files

Move the tsv files created by the carrot-cdm transform into this directory, e.g. in a subfolder mapped_data:

mapped_data/
test/
├── condition_occurrence.tsv
├── masked_person_ids.csv
├── observation.tsv
└── person.tsv

Get the name(s) of the tables

The default bclink table for "person" will be named person. However, you may want to create your own tables via the command line or via the GUI. As in the previous example, a person table could be created with the dataset id ds100394.

Clean the database [optional]

datasettool2 delete-all-rows ds100394 --database=bclink

Load the data

Use dataset_tool to load tsv files into the database (default name is bclink) as the user data (so you are able to view the upload jobs via the web GUI.

The flags --support & --bcqueue make sure the data file is uploaded as a job to the bclink batch queue.

Example:

dataset_tool --load \
             --table=ds100123 \
         --user=data
         --data_file=mapped_data/person.tsv \
         --support --bcqueue \
         bclink