--- license: other task_categories: - time-series-forecasting tags: - time-series - tsfile pretty_name: redset (TsFile format) --- # redset — TsFile 格式 本目录是 [`autogluon/fev_datasets`](https://huggingface.co/datasets/autogluon/fev_datasets) 中 **redset** 子集转换为 [Apache TsFile](https://tsfile.apache.org/) 格式的版本。 ## 来源与引用 - **原始来源**:https://github.com/amazon-science/redset/ - **论文/引用**:[[17]](https://www.amazon.science/publications/why-tpc-is-not-enough-an-analysis-of-the-amazon-redshift-fleet) - **统一格式合集**:[autogluon/fev_datasets](https://huggingface.co/datasets/autogluon/fev_datasets) > 本数据由外部来源转换为统一格式后再转为 TsFile。许可与引用以**原始来源**为准,我们不对原始数据主张任何权利。除非另有说明,数据仅供研究用途。 ## 数据统计 | 频率 | 序列数 | 中位长度 | 观测点数 | 动态列 | 静态列 | 文件 | |---|---|---|---|---|---|---| | 15T | 126 | 8,640 | 1,052,371 | 1 | 1 | `15T/15T_1..15T_2.tsfile`(2 片) | | 1H | 138 | 2,160 | 283,070 | 1 | 1 | `1H/1H.tsfile` | | 5T | 118 | 25,920 | 2,960,408 | 1 | 1 | `5T/5T_1..5T_3.tsfile`(3 片) | ## TsFile 存储模型 - 每条原始序列(`id`)→ 一个 **device**(TAG 维度)。 - 静态协变量列 → 也作 **TAG**(device 元数据):`subset`。 - 随时间变化的 target / 动态协变量 → **measurement**(FIELD)。 - `timestamp` → `Time`(INT64 毫秒)。 - 表名:redset_15T, redset_1H, redset_5T。 ### 列含义 | 列 | 角色 | TsFile 类型 | |---|---|---| | `Time` | Time(时间列) | INT64 | | `id` | TAG(device 维度) | STRING | | `subset` | TAG(device 维度) | STRING | | `target` | FIELD(measurement) | FLOAT | ## 转换说明 - 每行原始数据是一整条序列 `(id, timestamp[], 各 target[])`,纵向打平为长表后写入 TsFile。 - 数值类型按源列自适应:float32→FLOAT、float64→DOUBLE、整数→INT64、bool→BOOLEAN。 - 时间精度:毫秒(INT64)。 - 大表会被工具自动分片为 `<名>_1.tsfile`、`<名>_2.tsfile` …,同属一个逻辑表。 ## 读取示例 ```python from tsfile import TsFileReader reader = TsFileReader("15T/15T.tsfile") schemas = reader.get_all_table_schemas() # 表名:redset_15T;列见下方"列含义"。 ```