1 package de.dlr.shepard.data.timeseries.model.enums;
2
3 /**
4 * Possible format of CSV files. Currently, {@link CsvFormat#ROW} and
5 * {@link CsvFormat#COLUMN} are supported
6 */
7 public enum CsvFormat {
8 /**
9 * Row based CSV format. Each row consists of exactly one data point. The first 6 columns
10 * describe the timestamp and the timeseries 5-tuple, the last column contains the data value.
11 * Multiple timeseries are written completely one after each other, i.e. timestamps of
12 * data points from different timeseries are not sorted
13 */
14 ROW,
15 /**
16 * Column based CSV format. Each row contains all data points from all timeseries for one certain
17 * timestamp. If one timeseries does not contain data for a given timestamp, the column will be empty.
18 * This can lead to very sparse files.
19 * The timestamps are guaranteed to being ordered, but no assumptions can be made about the
20 * time-distance of two rows.
21 */
22 COLUMN,
23 }