Raw Data Export for Excel

The raw export delivers the measurements as recorded — any time window, any resolution (5 minutes up to a day), not bound to whole days — as a CSV you can open directly in Excel or LibreOffice. This guide walks through the four everyday cases: a single park's production, several parks side by side, individual inverters, and single strings.

While the template export is built for automated, repeatable reporting scripts, the raw export is made for manual analysis: pull a window into a spreadsheet, pivot it, chart it, done.

Prerequisites

Mirox Account with permission to read the park you export
API Token with the Reporting permission group — see API Token Usage. (In a logged-in browser, the shareable link works without a token.)

The Easy Path: Copy the Link from the UI

You never have to construct a URL by hand. Open Data export ▸ Raw data, pick your parks, metrics, window and resolution — the page shows a live API URL preview that mirrors your exact selection, including component choices. Copy it, append your token header in curl (or just open it in the browser), and you have the CSV.

The examples below show what those URLs look like, so you can adapt them in scripts or bookmark variations.

The Value Modes

Energy counters in the raw data are cumulative — they only ever grow. The value parameter decides what lands in your spreadsheet:

`value`	Meaning	Use it for
`plain` (default)	The counter as recorded, accumulating over the export window	Meter-reading style checks
`start_at_zero`	Each series shifted so it starts at 0	Comparing growth curves
`use_delta`	The change per interval — e.g. the energy produced in each 15-minute slot	Production analysis — this is what you usually want in Excel

Recipe 1: Single-Park Production

One week of a park's grid production at 15-minute resolution, as per-interval energy:

curl --compressed "https://service.mirox.io/api/v1/export/raw/query?metrics=raw_grid_energy_total&park=ABC123DEF456&start=2026-07-01T00:00:00Z&end=2026-07-08T00:00:00Z&step=15m&value=use_delta" \
  -H "Authorization: Bearer YOUR_API_TOKEN" -o production_week.csv

Date;Grid Production (Wh)
2026-07-01 00:00;0,00
2026-07-01 00:15;0,00
2026-07-01 06:15;1250,50
...

A full year at 15 minutes works the same way — set start/end a year apart (the in-app picker has This year and Last year presets for exactly this). That is the largest window per series the export allows; for multi-year windows switch to step=1h or 1d.

The raw metric catalog (ids, names, units) is at GET /v1/metrics/raw; up to 20 metrics per request, comma-separated.

Recipe 2: Multi-Park Comparison

Several parks in one file, one column per park (multi_park_agg=split):

curl --compressed "https://service.mirox.io/api/v1/export/raw/query?metrics=raw_grid_energy_total&park=ABC123DEF456,DEF456ABC123&start=2026-07-01T00:00:00Z&end=2026-08-01T00:00:00Z&step=1h&value=use_delta&multi_park_agg=split" \
  -H "Authorization: Bearer YOUR_API_TOKEN" -o parks_compare.csv

Date;Sunnyside - Grid Production (Wh);Riverbend - Grid Production (Wh)
2026-07-01 00:00;0,00;0,00
...

multi_park_agg=merge (the default) combines the parks into one column per metric instead — summed for energy/power, averaged for temperatures and ratios.
portfolio=<uid> scopes to all parks of a portfolio without listing them.
A split export is capped at 50 columns (metrics × parks).

Recipe 3: Individual Inverters

The component export drills into one park's components. Metric ids carry the comp_raw_ prefix (catalog: GET /v1/metrics/components?component_type=inverter):

curl --compressed "https://service.mirox.io/api/v1/export/raw/component/query?metrics=comp_raw_inverter_energy_ac&park=ABC123DEF456&component_type=inverter&start=2026-07-01T00:00:00Z&end=2026-07-08T00:00:00Z&step=15m&value=use_delta" \
  -H "Authorization: Bearer YOUR_API_TOKEN" -o inverters.csv

Date;WR 01.01 - Inverter Production (Wh);WR 01.02 - Inverter Production (Wh);...
2026-07-01 06:15;620,25;618,75;...

Without a components selection, the export pages through all inverters of the park (page/limit, up to 50 per page).
To pick specific ones, pass components=<id1>,<id2>,... — the ids appear in the UI's component picker (and its copied URL), or in the export response's components list.
component_type=gak works identically for combiner boxes.
multi_component_agg=merge collapses the selection into one column (sum for energy, average for voltages/temperatures).

Recipe 4: Single Strings

Same endpoint with component_type=string — ideal for checking one suspicious string against its siblings:

curl --compressed "https://service.mirox.io/api/v1/export/raw/component/query?metrics=comp_raw_string_energy&park=ABC123DEF456&component_type=string&components=1084279778457255938&start=2026-07-01T00:00:00Z&end=2026-07-08T00:00:00Z&step=15m&value=use_delta" \
  -H "Authorization: Bearer YOUR_API_TOKEN" -o string.csv

The easiest way to get string ids is the in-app picker on the Raw data tab (detail level String): select the strings, then copy the previewed URL — it already carries the components list.

Opening the CSV in Excel

Separators: the defaults (; field, , decimal) open directly in German-locale Excel. For US/UK locales pass separator_csv=,&separator_decimal=. — the two must differ.
Dates: sub-daily steps include the time of day; datetime_format accepts a strftime pattern (e.g. %d.%m.%Y) if your locale needs it.
Encoding: files are UTF-8; use Data ▸ From Text/CSV if umlauts look wrong.
Localized headers: language=de (or es, fr, it, pt) translates the column names.
Size: large responses are gzip-compressed on the wire when the client accepts it — hence --compressed in the examples; browsers handle this automatically.

Limits

Limit	Value
Metrics per request	20
Data points per series (park level)	180,000 — 5 years at 15-minute resolution fits
Total samples (park level)	1,800,000 (series × data points per series)
Data points per series (component level)	36,000 — a full year at 15-minute resolution fits
Park `split` columns	50 (metrics × parks)
Explicitly selected components	50
Component `split` columns	100 (metrics × components)

A request over budget is rejected with 400 and a message naming the limit — increase the step, shorten the window, or reduce the selection. The full reference lives in the Metrics Export API Guide.

External Report Generator — the template-export counterpart: automated Python reporting
Metrics Export API — full parameter and limits reference for all export paths
API Token Usage — create the token the curl examples use
MiroxQL — build derived metrics for the aggregated exports