File Destination Plugin

Example #

kind: destination
spec:
  name: "file"
  path: "cloudquery/file"
  registry: "cloudquery"
  version: "v5.4.3"
  write_mode: "append"
  # Learn more about the configuration options at https://cql.ink/file_destination
  spec:
    path: "path/to/files/{{TABLE}}/{{UUID}}.{{FORMAT}}"
    format: "parquet" # options: parquet, json, csv
    # Optional parameters
    # format_spec:
      # CSV specific parameters:
      # delimiter: ","
      # skip_header: false
      # Parquet specific parameters:
      # version: "v2Latest"
      # root_repetition: "repeated"
      # max_row_group_length: 134217728 # 128 * 1024 * 1024
    # compression: "" # options: gzip
    # no_rotate: false
    # batch_size: 10000
    # batch_size_bytes: 52428800 # 50 MiB
    # batch_timeout: 30s

File Spec #

• path (string) (required)

  Path template string that determines where files will be written, for example path/to/files/{{TABLE}}/{{UUID}}.parquet.

  The path supports the following placeholder variables:

  • {{TABLE}} will be replaced with the table name
  • {{FORMAT}} will be replaced with the file format, such as csv, json or parquet. If compression is enabled, the format will be csv.gz, json.gz etc.
  • {{UUID}} will be replaced with a random UUID to uniquely identify each file
  • {{YEAR}} will be replaced with the current year in YYYY format
  • {{MONTH}} will be replaced with the current month in MM format
  • {{DAY}} will be replaced with the current day in DD format
  • {{HOUR}} will be replaced with the current hour in HH format
  • {{MINUTE}} will be replaced with the current minute in mm format

  Note that timestamps are in UTC and will be the current time at the time the file is written, not when the sync started.
• format (string) (required)

  Format of the output file. Supported values are csv, json and parquet.
• format_spec (format_spec) (optional)

  Optional parameters to change the format of the file.
• no_rotate (boolean) (optional) (default: false)

  If set to true, the plugin will write to one file per table. Otherwise, for every batch a new file will be created with a different .<UUID> suffix.
• compression (string) (optional) (default: "")

  Compression algorithm to use. Supported values are "" and gzip. Not supported for parquet format.
• batch_size (integer) (optional) (default: 10000)

  Number of records to write before starting a new file.
• batch_size_bytes (integer) (optional) (default: 52428800 (50 MiB))

  Number of bytes (as Arrow buffer size) to write before starting a new file.
• batch_timeout (duration) (optional) (default: 30s (30 seconds))

  Maximum interval between batch writes.

format_spec #

CSV

• delimiter (string) (optional) (default: ,)

  Delimiter to use in the CSV file.
• skip_header (boolean) (optional) (default: false)

  If set to true, the CSV file will not contain a header row as the first row.

JSON

Parquet

• version (string) (optional) (default: v2Latest)

  Parquet format version to use. Supported values are v1.0, v2.4, v2.6 and v2Latest. v2Latest is an alias for the latest version available in the Parquet library which is currently v2.6.

  Useful when the reader consuming the Parquet files does not support the latest version.
• root_repetition (string) (optional) (default: repeated)

  Repetition option to use for the root node. Supported values are undefined, required, optional and repeated.

  Some Parquet readers require a specific root repetition option to be able to read the file. For example, importing Parquet files into Snowflake requires the root repetition to be undefined.
• max_row_group_length (integer) (optional) (default: 134217728 (= 128 * 1024 * 1024))

  The maximum number of rows in a single row group. Use a lower number to reduce memory usage when reading the Parquet files, and a higher number to increase the efficiency of reading the Parquet files.