Running an aggregation from code

An aggregation is not only for the portal. Northgate Energy's risk team wants the same delivery check in a Python notebook, and the operations team calls it from a script before they release the evening's curves. Code in the workspace toolbar opens the aggregation as code you can copy or download.

---

The languages

Language	What it does	When to use it
ODSL	The aggregate command with the stages written in, and the test dates filled in	A one-off run in a script or the ODSL console
ODSL (saved script)	Imports the saved script and calls its function with dates	Scripts and processes that should always run the saved version. Save the aggregation first.
curl	One request to the REST API, with the pipeline in the _aggregate parameter	Trying the REST API, or a shell script
Python	A function that calls the REST API with the requests package and returns the rows	Notebooks and Python services
JavaScript	The same with fetch, for Node 18 or later	Node services and tools

Copy puts the code on the clipboard and Download saves it as a file.

---

Signing in

For curl, Python and JavaScript, choose Sign in with:

Option	What the code sends	Set these environment variables
An API key	The headers x-odsl-email and x-odsl-key	ODSL_EMAIL, ODSL_KEY
An access token	A bearer token	ODSL_TOKEN

The code reads the values from environment variables, so they are never written into it. When the portal is not on the production environment, the code also sends the x-odsl-environment header.

---

ODSL

Once the aggregation is saved, this is the way to call it from ODSL. It imports the saved script, so the script always runs the latest saved pipeline:

// Dataset deliveries for the day: calls the saved aggregation script
import ${script:"aggregator\dataset_deliveries"}

ondate = Date("2026-10-15")
fromDate = Date("2026-10-01")
toDate = Date("2026-10-15")
result = aggregate_dataset_deliveries(ondate, fromDate, toDate)
print result

---

curl

The curl code has the test dates written into the pipeline:

# Dataset deliveries for the day on Dataset deliveries (common)
# The date placeholders are filled with the test values: ondate 2026-10-15
# Set ODSL_EMAIL and ODSL_KEY to your email address and API key
curl -G "https://api.opendatadsl.com/api/dataset_delivery/v1/common" \
  -H "x-odsl-email: $ODSL_EMAIL" \
  -H "x-odsl-key: $ODSL_KEY" \
  --data-urlencode "_limit=-1" \
  --data-urlencode '_aggregate=[{"$match":{"ondate":"2026-10-15"}},{"$project":{"_id":0,"dataset":"$dsid","status":1,"loaded":"$timings.loaded","score":1}},{"$sort":{"score":1,"dataset":1}}]'

---

Python and JavaScript

The Python and JavaScript code keep the placeholders in the pipeline and fill them from PARAMS, so you can run them for other dates. In Python:

rows = run_aggregation({"ondate": "2026-10-16"})

and in JavaScript:

const rows = await runAggregation({ondate: "2026-10-16"});

When the aggregation uses report properties, PARAMS holds them too, with their test values, and the curl code has the test values written in. ODSL (saved script) calls the function without a report, so the report properties take their test values.

note

The code runs the pipeline as it is in the workspace when you open Code, not the saved script. If you change the aggregation later, copy the code again, or use ODSL (saved script).

---

Next

Creating a report: store the result for each date and show it in the house style.