ECB Data Portal / euro reference rates

Name: ECB Data Portal / euro reference rates
Creator: ECB Data Portal / euro reference rates
License: ecb-attrib-nomodify

Open · redistributedecb-attrib-nomodifyfxrates

Redistributable. This source is served as canonical Parquet under the license below, with attribution and provenance preserved.

Coverage

Series catalogued

Temporal coverage

1994-01-01 – 2026-06-03

Frequencies

Daily, Monthly

Distinct geographies

Licensing & provenance

License: ecb-attrib-nomodify
Required attribution: Source: ECB -- available free at the ECB; values not modified
Redistribution: Permitted (served here)
Commercial use: Yes
Modification: Restricted

How to cite

European Central Bank, Statistical Data Warehouse. Retrieved from https://data.ecb.europa.eu. Compiled and redistributed by the Elkassabgi Data Library.

Processing: Retrieved from the official source, normalized to a long {series_key, obs_date, value} schema (period-start dates), de-duplicated, and stored as zstd Parquet. Compiled and redistributed by the Elkassabgi Data Library.

Access & mirrors

Download: Select & download ecb series as CSV →
API: Get a free API key, then GET /v1/series/<id>.csv
Canonical landing: https://econdatalibrary.com/ecb.html
Update cadence: daily
Update strategy: sdmx delta
Storage layout: grouped parquet, many files (~542 on disk): <AGENCY>__<FLOW>.parquet for whole pulls, <AGENCY>__<FLOW>__<FREQ>.parquet for frequency chunks, <AGENCY>__<FLOW>__<FREQ>__<VAL>[__<YEAR>].parquet for sub/decade/year splits. Schema series_key(str, ECB pre-joined key e.g. EXR.D.USD.EUR.SP00.A), obs_date(date32), value(float64), freq(str). Sidecars _manifest.json, _catalog.json. Writers stream to .part then atomic os.replace.

Full description

FULL-COVERAGE grouped ingest of the entire ECB Data Portal (SDMX 2.1). Source: ECB Data Portal (data-api.ecb.europa.eu). license_id = ecb-attrib-nomodify (from configs/sources.yaml: "Source: ECB -- available free at the ECB; values not modified"). We do NOT modify values; the Clean tier stores the raw observation. Strategy (verified live against data-api.ecb.europa.eu, June 2026): * Catalog = EVERY dataflow from /service/dataflow (213 flows across agencies ECB, ECB.DISS, ESTAT, EUROSTAT, IMF -- ALL hosted at the ECB endpoint). * Data pulled as SDMX-CSV (?format=csvdata) with &detail=dataonly, which strips the ~25 bulky attribute columns (TITLE, OBS_COM, COLLECTION, ...) WITHOUT changing the observation count and is ~10-30x faster to transfer. The CSV header is then: KEY, <dim1..dimN>, TIME_PERIOD, OBS_VALUE. * SERIES KEY = the CSV `KEY` column, which the ECB already emits pre-joined (e.g. "EXR.D.USD.EUR.SP00.A") -> stored verbatim as column `series_key`. FREQ (when present as a column) -> stored as `freq`. * Most flows pull whole in one request. The handful of GIANTS (EXR ~197MB, MIR ~124MB, YC ~11M rows, BSI, ICP/HICP, STS, SEC, IRS, FM, BOP, RPP, ...) are FREQUENCY-CHUNKED via the SDMX positional path key (the DSD dimension order, FREQ value + dots for the rest, e.g. EXR key "D....."). If a single frequency chunk is still huge/fails, it is SUB-SPLIT by the next splitting dimension's codelist values (e.g. YC daily -> by INSTRUMENT_FM curve type). Each chunk is an independent, restart-safe request that lands in its own parquet part, so a process kill costs at most one chunk. OUTPUT: grouped Parquet under data/clean_full/ecb/ -> <AGENCY>__<FLOW>.parquet (whole-flow pulls) <AGENCY>__<FLOW>__<FREQ>.parquet (frequency chunks) <AGENCY>__<FLOW>__<FREQ>__<VAL>.parquet (sub-split chunks) columns: series_key (str), obs_date (date32), value (float64), freq (str). A sidecar _manifest.json records per-flow stats (NOT catalog.db, which is off-limits). Memory is bounded: every HTTP body streams to a temp .csv.gz on disk, is parsed row-by-row, and flushed to the Parquet writer in RecordBatches; no flow is held whole in RAM. The distinct-series set per file is capped to bound memory. Usage: python jobs/ingest_ecb.py --list # enumerate catalog, no data python jobs/ingest_ecb.py --only EXR,YC # just these flows python jobs/ingest_ecb.py --workers 4 # FULL run, 4 concurrent flows python jobs/ingest_ecb.py --report # re-read parquet, print true totals python jobs/ingest_ecb.py # FULL run (default workers=4)

Structured metadata

schema.org/Dataset (Google Dataset Search)

{
  "@context": "https://schema.org/",
  "@type": "Dataset",
  "name": "ECB Data Portal / euro reference rates",
  "description": "FULL-COVERAGE grouped ingest of the entire ECB Data Portal (SDMX 2.1). Source: ECB Data Portal (data-api.ecb.europa.eu). license_id = ecb-attrib-nomodify (from configs/sources.yaml: \"Source: ECB -- available free at the ECB; values not modified\"). We do NOT modify values; the Clean tier stores the…",
  "url": "https://econdatalibrary.com/ecb.html",
  "identifier": "ecb",
  "isAccessibleForFree": true,
  "publisher": {
    "@type": "Organization",
    "name": "Econ Data Library",
    "url": "https://econdatalibrary.com"
  },
  "creator": {
    "@type": "Organization",
    "name": "ECB Data Portal / euro reference rates"
  },
  "license": "ecb-attrib-nomodify",
  "keywords": [
    "fx",
    "rates",
    "ecb"
  ],
  "temporalCoverage": "1994-01-01/2026-06-03",
  "repeatFrequency": [
    "P1D",
    "P1M"
  ],
  "creditText": "Source: ECB -- available free at the ECB; values not modified",
  "citation": "European Central Bank, Statistical Data Warehouse. Retrieved from https://data.ecb.europa.eu. Compiled and redistributed by the Elkassabgi Data Library.",
  "sameAs": [
    "https://huggingface.co/datasets/econdatalibrary-ecb",
    "https://zenodo.org/communities/econdatalibrary"
  ],
  "distribution": [
    {
      "@type": "DataDownload",
      "encodingFormat": "text/csv",
      "contentUrl": "https://econdatalibrary.com/download.html?source=ecb"
    }
  ]
}

Croissant (ML-ready, schema.org JSON-LD)

{
  "@context": {
    "@language": "en",
    "@vocab": "https://schema.org/",
    "cr": "http://mlcommons.org/croissant/",
    "data": {
      "@id": "cr:data",
      "@type": "@json"
    },
    "dataType": {
      "@id": "cr:dataType",
      "@type": "@vocab"
    },
    "sc": "https://schema.org/",
    "conformsTo": "dct:conformsTo",
    "dct": "http://purl.org/dc/terms/"
  },
  "@type": "sc:Dataset",
  "conformsTo": "http://mlcommons.org/croissant/1.0",
  "name": "ecb",
  "description": "FULL-COVERAGE grouped ingest of the entire ECB Data Portal (SDMX 2.1). Source: ECB Data Portal (data-api.ecb.europa.eu). license_id = ecb-attrib-nomodify (from configs/sources.yaml: \"Source: ECB -- available free at the ECB; values not modified\"). We do NOT modify values; the Clean tier stores the…",
  "url": "https://econdatalibrary.com/ecb.html",
  "creator": {
    "@type": "sc:Organization",
    "name": "ECB Data Portal / euro reference rates"
  },
  "publisher": {
    "@type": "sc:Organization",
    "name": "Econ Data Library",
    "url": "https://econdatalibrary.com"
  },
  "license": "ecb-attrib-nomodify",
  "keywords": [
    "fx",
    "rates"
  ],
  "temporalCoverage": "1994-01-01/2026-06-03",
  "citation": "European Central Bank, Statistical Data Warehouse. Retrieved from https://data.ecb.europa.eu. Compiled and redistributed by the Elkassabgi Data Library.",
  "distribution": [
    {
      "@type": "cr:FileObject",
      "@id": "ecb-csv",
      "name": "ecb-csv",
      "description": "Per-series CSV, downloadable with a free API key.",
      "contentUrl": "https://econdatalibrary.com/download.html?source=ecb",
      "encodingFormat": "text/csv"
    }
  ]
}

Part of the Econ Data Library catalog · metadata generated 2026-07-14 from the central registry · browse all datasets