op-export_ts

This operator allow to export a dataset to a set of CSV files (see format below following a defined path pattern.

The directory tree can be customized to define the name of the folders, subfolders and filenames based on timeseries metadata.

Overview

Input and parameters

The operator takes 1 input:

• DSname: The dataset name (string) to extract timeseries from

It also takes 1 parameter from the user:

• Pattern: The path pattern used to build the directory tree (see pattern format below).

Outputs

The operator has 1 Output:

• Path: String indicating where the CSV files are located on container disk relative to TSDATA environment variable

Details

Pattern format

The path pattern is a string containing placeholders for specific metadata.

Placeholders are defined using curly brackets surrounding the keyword.

As an example, you could create a directory tree where:

• the first folder level is the number of points of the timeseries
• the second level is the maximum value of the timeseries
• the filename is composed of a city and rank metadata

The corresponding pattern could be: {qual_nb_points}/{max_value}/{city}_{rank}.csv

Reserved keywords are:

• {fid} for the functional identifier (the displayed name of the timeseries)
• {DSname} for the name of the dataset (specified as input)

CSV format

The CSV format is composed of 2 columns delimited by the character ; Columns:

• Date (ISO-8601) (eg. 2018-01-01 12:34:56.789)
• Value (float)

Limitations

• If the given path pattern produces 2 or more identical file path, the export is aborted.
• If the given path pattern can't be produced due to unknown metadata, the following pattern is used as fallback {fid}.csv