Populate DHIS2 Tracker with individual data records read from a number of CSV files.

This is a node.js application. Refer to http://nodejs.org/ for instructions on how install nodejs and NPM for your operating system.

Install the package globally using NPM.

npm install -g dhis2-tracker-populator

The CSV files must follow a naming convention of programID.programStageID.trackedEntityID.csv. The values that make up the filename will be used when making requests to the API. These are the uids of the program, stage and trackedentity which can be found using api/resource calls, for example:

Program: https://apps.dhis2.org/demo/api/programs/IpHINAT79UW (shows attributes)

Stage: https://apps.dhis2.org/demo/api/programStages/A03MvHHogjR (shows data elements)

The CSV files must contain the following columns:

• orgUnit : the uid of the organisation unit
• programDate : yyyy-mm-dd
• eventDate : yyyy-mm-dd

Any number of additional attributes or data elements can be specified with column headers of the following format:

• Attributes: The uid of the attribute prepended with the string A|.
• Data elements: The uid of the data element prepended with the string DE|.
• Each line in the CSV corresponds to an event.
• If the tracked entity exists already (there is a unique attribute), other atrributes will not be updated
• If the tracked entity is already enrolled into the program, the application will continue on to adding the event
• There is no restriction on the number of events added to a program for a particular tracked entity, and multiple identical events can be added (for example if the same csv file is operated on more than once)
• The application assumes that there is only one unique attribute per tracked entity (this is determined by the application at run time from the DHIS2 resources api)
• The csv file can be generated from a sql query to a database, from an excel spreadsheet, or any other source, as ong as it is added to the csv directory in the proper format (this must actually be comma separated)

e.g. file: IpHINAT79UW.A03MvHHogjR.cyl5vuJ5ETQ.csv

orgUnit,programDate,eventDate,A|dv3nChNSIxy,A|hwlRTFIFSUq,DE|UXz7xuGCEhU
sY1WN6LjmAx,2014-08-11,2014-08-11,babyfirstname,babylastname,3000

Run the populator from the command line.

Usage: dhis2-tracker-populator [OPTIONS] URL

Options:
  -c, --csv        Path to the directory containing the csv files
  -d, --done       Path to the directory in which to place the done files
  -f, --fail       Path to the directory in which to place the failed files
  -t, --threshold  The minimum number of days between duplicate events

  --help      Show this help
  --version   Print the version and exit

NB: The folders specified in the options above need to be created manually

The threshold option specifies the number of days between events for the same program, stage and tracked entity instance which will cause them to be considered duplicates. For example, if set to 0 events on the same day will be considered duplicates. The default value of -1 disables duplicate checking.

The default options are:

• csv: ./csv/
• done: ./csvdone/
• fail: ./csvfail/
• threshold: -1

Example:

dhis2-tracker-populator https://username:[email protected]/demo

This will read data records from the CSV files found in the csv directory and make requests to the API for the given URL to populate the tracker. If an error occurs while processing a CSV file it will be moved into the csvfail directory, otherwise it will be moved to the csvdone directory on completion.

Best practice is to use ./bin/split.js to split a multi-line csv into individual files. This will give granular control and a view of exactly which files have been processed and which did not get fully added to the program.

To remove ALL tracker data from your DHIS2 database run the following sql commands in order:

delete from trackedentitydatavalue;
delete from programstageinstance;
delete from programinstance;
delete from trackedentityaudit;
delete from trackedentityattributevalue;
delete from trackedentityinstance;

WARNING: This cannot be undone