Skip to content
HALOWERK

Formatwerk

csv parquet convert

Runs DuckDB 2.0 in memory with one thread and 384 MB: the input is written to a private temporary folder, DuckDB may only read and write inside that folder (allowed_directories), external access is disabled and the configuration is locked before any customer data is read; nothing is kept after the call. CSV types are auto-detected over all rows (DuckDB sniffer) and can be overridden per column; Parquet output uses ZSTD, Snappy, GZIP or no compression. The returned schema lists physical,.

POST https://format.halowerk.com/v1/csv-parquet-convert

Inputs

Fields of the published input schema
FieldTypeRequiredDescription
directionstringYesConversion direction.
csvstringNoCSV text (for csv_to_parquet).
delimiterstringNoCSV delimiter; auto-detected if omitted.
headerbooleanNoWhether the CSV has (csv_to_parquet) or gets (parquet_to_csv) a header row.
column_typesobjectNoOptional DuckDB type per column name, e.g. {"amount": "DECIMAL(12,2)"}.
compressionstringNoParquet compression codec.
parquet_base64stringNoParquet file as base64 (for parquet_to_csv).
output_delimiterstringNoDelimiter of the CSV output.
Show full input schema
{
  "type": "object",
  "additionalProperties": false,
  "required": [
    "direction"
  ],
  "properties": {
    "direction": {
      "type": "string",
      "enum": [
        "csv_to_parquet",
        "parquet_to_csv"
      ],
      "description": "Conversion direction."
    },
    "csv": {
      "type": "string",
      "minLength": 1,
      "maxLength": 5000000,
      "description": "CSV text (for csv_to_parquet)."
    },
    "delimiter": {
      "type": "string",
      "enum": [
        ",",
        ";",
        "\t",
        "|"
      ],
      "description": "CSV delimiter; auto-detected if omitted."
    },
    "header": {
      "type": "boolean",
      "default": true,
      "description": "Whether the CSV has (csv_to_parquet) or gets (parquet_to_csv) a header row."
    },
    "column_types": {
      "type": "object",
      "maxProperties": 200,
      "additionalProperties": {
        "type": "string",
        "pattern": "^(BOOLEAN|TINYINT|SMALLINT|INTEGER|BIGINT|HUGEINT|UTINYINT|USMALLINT|UINTEGER|UBIGINT|FLOAT|DOUBLE|DATE|TIME|TIMESTAMP|TIMESTAMPTZ|VARCHAR|UUID|INTERVAL|DECIMAL\\(\\d{1,2},\\d{1,2}\\))$"
      },
      "description": "Optional DuckDB type per column name, e.g. {\"amount\": \"DECIMAL(12,2)\"}."
    },
    "compression": {
      "type": "string",
      "enum": [
        "zstd",
        "snappy",
        "gzip",
        "uncompressed"
      ],
      "default": "zstd",
      "description": "Parquet compression codec."
    },
    "parquet_base64": {
      "type": "string",
      "minLength": 16,
      "maxLength": 8000000,
      "description": "Parquet file as base64 (for parquet_to_csv)."
    },
    "output_delimiter": {
      "type": "string",
      "enum": [
        ",",
        ";",
        "\t",
        "|"
      ],
      "default": ",",
      "description": "Delimiter of the CSV output."
    }
  }
}

Open original schema

Published input example

{
  "direction": "csv_to_parquet",
  "csv": "id,amount,name\n1,19.90,Ada\n2,5.00,Grace\n",
  "column_types": {
    "amount": "DECIMAL(10,2)"
  }
}

Result & documentation

Machine-readable documentation of this hub

Source of this page: the published HALOWERK catalog. Updated: 2026-11-17 UTC.

More endpoints in this hub

All endpoints in this hub โ†’