Timeseries
A timeseries is a list of values against times, aligned to a calendar. You save a timeseries as a property of a master data object and read it with the Data service using its data id, objectid:property, for example NGE.POWER.DE:SPOT.
There are five kinds:
_type | What it is |
|---|---|
VarTimeSeries | Values you load |
VarSmartTimeSeries | Calculated from other timeseries with an expression |
VarEventTimeSeries | Built from the events of an event list |
VarMetricTimeSeries | Built from platform usage metrics |
VarComposedTimeSeries | An element of a composition |
Smart, event, metric and composed timeseries are definitions: their values are worked out when they are read (or built), so saving one with a data field is refused.
TimeSeries
| Field | Type | Required | What it is |
|---|---|---|---|
_type | String | Yes | VarTimeSeries |
_id | String | No | The data id within the object. When saved in an object, the property name is used |
calendar | String | No | The calendar code, for example DAILY, BUSINESS, #HENG, #PT1H. Defaults to SPARSE (no fixed interval) |
start | Date | No | The time of the first value. Needed only when data is a list of plain numbers |
data | Array | No | The values, see Values |
name | String | No | A name |
description | String | No | A description |
currency | String | No | The currency code, for example EUR |
units | String | No | The units code, for example MWH, see the Unit service |
timezone | String | No | The timezone of the times, for example Europe/Berlin |
observed | String | No | How a value relates to its period, used when converting to another calendar: end (the default), beginning, averaged, summed, high, low, delta or none |
tenor | String | No | A tenor period code, for timeseries taken from a curve |
source | String | No | Where the data comes from |
dataType | String | No | The type of the values, defaults to any |
valueType | String | No | TRACKED (the default): values are objects with a status. BASIC: values are written with only time and value |
precision | Object | No | Decimal places and rounding, see Numbers |
scaling | Integer | Output | The scaling applied when the timeseries was read |
properties | Object | No | Your own properties |
_status | Object | No | The status of the whole timeseries, see Status |
_checks | Array | Output | The quality checks recorded on the timeseries |
_identity | Object | No | The codes other systems use for the timeseries, by provider, see Identities |
_links | Object | No | Links |
_objid, _dataid, _objtype, _subtype | String | Output | Set by the platform |
Values
Each value in data is an object:
| Field | Type | What it is |
|---|---|---|
time | Date or DateTime | The time of the value |
value | Number | The value, or "NaN" for a missing value |
status | Object | Optional: the status of the value |
Responses for a tracked timeseries also have "_type": "TimeValue", "_links": {} and "allowLinks": true on each value; you do not need to send them. Fields with no value (for example a timeseries with no currency) are left out of the JSON.
For a calendar with a fixed interval you can also send data as a list of plain numbers, with the time of the first one in start. Each number takes the next time of the calendar.
{
"_type": "VarTimeSeries",
"calendar": "BUSINESS",
"start": "2026-11-16",
"data": [41.85, 42.10, 41.95, 42.60, 43.05]
}
A SPARSE timeseries must use value objects.
Example
{
"_id": "NGE.GAS.TTF:SPOT",
"_type": "VarTimeSeries",
"start": "2026-11-16",
"calendar": "BUSINESS",
"observed": "end",
"currency": "EUR",
"units": "MWH",
"name": "TTF day-ahead",
"description": "Northgate TTF day-ahead index",
"precision": {"roundingMode": "HALF_UP", "scale": 6, "useDefault": true, "store": false},
"valueType": "TRACKED",
"dataType": "any",
"properties": {},
"_links": {},
"_status": {},
"_checks": [],
"_dataid": "SPOT",
"_objid": "NGE.GAS.TTF",
"_objtype": "#Object",
"_identity": {"ETRM": "GAS_TTF_DA", "RISK": "TTF-DA-IDX"},
"data": [
{"time": "2026-11-16", "value": 41.85, "status": {}, "_type": "TimeValue", "_links": {}, "allowLinks": true},
{"time": "2026-11-17", "value": 42.10, "status": {}, "_type": "TimeValue", "_links": {}, "allowLinks": true},
{"time": "2026-11-18", "value": "NaN", "status": {"quality": "FAILED:missing"}, "_type": "TimeValue", "_links": {}, "allowLinks": true}
],
"_oid": "NGE.GAS.TTF:SPOT",
"_version": 12,
"_user": "priya.shah@northgate-energy.com",
"_timestamp": "2026-11-18T18:05:44.120",
"_tag": []
}
Smart timeseries
A smart timeseries calculates its values from a base timeseries (and optionally other inputs) with an expression. It has the descriptive fields of a timeseries (name, description, currency, units, timezone, precision, properties) and:
| Field | Type | Required | What it is |
|---|---|---|---|
_type | String | Yes | VarSmartTimeSeries |
baseTimeSeries | Reference or String | Yes | The timeseries it is built on: a reference, or a data id |
expression | String | Yes | The ODSL expression, for example BASE * 1.1 |
script | String | No | A script whose functions the expression can use |
calendar | String | No | Convert the result to this calendar |
properties | Object | No | Other inputs. A string that is a data id (it contains a colon) is turned into a reference to that data |
{
"_id": "SPOT_GBP",
"_type": "VarSmartTimeSeries",
"name": "TTF day-ahead in GBP",
"currency": "GBP",
"units": "THM",
"baseTimeSeries": {
"_type": "VarReference",
"_service": "data",
"_source": "private",
"_oid": "NGE.GAS.TTF:SPOT"
},
"expression": "BASE * FX",
"properties": {
"FX": {
"_type": "VarReference",
"_service": "data",
"_source": "public",
"_oid": "#ECB_FX.EURGBP:SPOT"
}
}
}
See Smart Data for building smart timeseries and curves.
Event timeseries
An event timeseries takes one property of the events in an event list as its values.
| Field | Type | Required | What it is |
|---|---|---|---|
_type | String | Yes | VarEventTimeSeries |
event | String | Yes | The event list, as objectid:eventlist |
property | String | Yes | The event property used as the value |
filter | String | No | A filter on the events, for example {'side': 'ASK'} |
index | String | No | The event date used as the time: eventtime, eventstart or eventend |
calendar | String | No | The calendar, defaults to SPARSE |
{
"_id": "FILL_PRICE",
"_type": "VarEventTimeSeries",
"event": "NGE.TRADES:FILLS",
"property": "price",
"filter": "{'product': 'TTF'}"
}
Metric timeseries
A metric timeseries totals or averages platform usage metrics into a timeseries.
| Field | Type | Required | What it is |
|---|---|---|---|
_type | String | Yes | VarMetricTimeSeries |
provider | String | Yes | The metric provider, for example ODSLRequest or ODSLProcess |
metric | String | Yes | The metric name, for example ReadRequests |
filter | String | No | A filter on the metrics |
calendar | String | No | The calendar to aggregate to, for example #PT1H |
observed | String | No | How to aggregate: summed, averaged, high, low, beginning or end |
{
"_id": "PROCESS_RUNS_HOURLY",
"_type": "VarMetricTimeSeries",
"provider": "ODSLProcess",
"metric": "RunRequests",
"calendar": "#PT1H",
"observed": "summed"
}
See the Metric service for the providers and metrics.
Composed timeseries
A composed timeseries is one element of a composition report, saved so it can be read like any other timeseries.
| Field | Type | Required | What it is |
|---|---|---|---|
_type | String | Yes | VarComposedTimeSeries |
composition | String | Yes | The id of the composition |
element | String | Yes | The element (column) of the composition |
name, description | String | No | Descriptive fields |
properties | Object | No | Your own properties |
{
"_id": "SPARK_SPREAD",
"_type": "VarComposedTimeSeries",
"composition": "NGE_SPARK_SPREADS",
"element": "spread",
"name": "German clean spark spread"
}