VGalaxies666 commited on
Commit
8bb7b3e
·
verified ·
1 Parent(s): c7fac0c

Copy rossmann from smilegeng/fev_datasets

Browse files
Files changed (5) hide show
  1. .gitattributes +2 -0
  2. 1D/1D_1.tsfile +3 -0
  3. 1D/1D_2.tsfile +0 -0
  4. 1W/1W.tsfile +3 -0
  5. README.md +79 -0
.gitattributes CHANGED
@@ -58,3 +58,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
58
  # Video files - compressed
59
  *.mp4 filter=lfs diff=lfs merge=lfs -text
60
  *.webm filter=lfs diff=lfs merge=lfs -text
 
 
 
58
  # Video files - compressed
59
  *.mp4 filter=lfs diff=lfs merge=lfs -text
60
  *.webm filter=lfs diff=lfs merge=lfs -text
61
+ 1D/1D_1.tsfile filter=lfs diff=lfs merge=lfs -text
62
+ 1W/1W.tsfile filter=lfs diff=lfs merge=lfs -text
1D/1D_1.tsfile ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:63c101ecc39689fcb09b562e71aec4d467b8ce5503aa54a8971475c0a477670b
3
+ size 7099229
1D/1D_2.tsfile ADDED
Binary file (13 kB). View file
 
1W/1W.tsfile ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fe0f2a353a4a52b9e436b25af34731b923f4eeb312971f961309f27a05e5391d
3
+ size 1897682
README.md ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ task_categories:
4
+ - time-series-forecasting
5
+ tags:
6
+ - time-series
7
+ - tsfile
8
+ pretty_name: rossmann (TsFile format)
9
+ ---
10
+
11
+ # rossmann — TsFile 格式
12
+
13
+ 本目录是 [`autogluon/fev_datasets`](https://huggingface.co/datasets/autogluon/fev_datasets) 中 **rossmann** 子集转换为 [Apache TsFile](https://tsfile.apache.org/) 格式的版本。
14
+
15
+ ## 来源与引用
16
+
17
+ - **原始来源**:https://www.kaggle.com/competitions/rossmann-store-sales
18
+ - **论文/引用**:[[21]](https://www.kaggle.com/competitions/rossmann-store-sales/overview/citation)
19
+ - **统一格式合集**:[autogluon/fev_datasets](https://huggingface.co/datasets/autogluon/fev_datasets)
20
+
21
+ > 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。
22
+
23
+ ## 数据统计
24
+
25
+ | 频率 | 序列数 | 中位长度 | 观测点数 | 动态列 | 静态列 | 文件 |
26
+ |---|---|---|---|---|---|---|
27
+ | 1D | 1,115 | 942 | 7,352,310 | 7 | 10 | `1D/1D_1..1D_2.tsfile`(2 片) |
28
+ | 1W | 1,115 | 133 | 889,770 | 6 | 10 | `1W/1W.tsfile` |
29
+
30
+ ## TsFile 存储模型
31
+
32
+ - 每条原始序列(`id`)→ 一个 **device**(TAG 维度)。
33
+ - 静态协变量列 → 也作 **TAG**(device 元数据):`Store, StoreType, Assortment, CompetitionDistance, CompetitionOpenSinceMonth, CompetitionOpenSinceYear, Promo2, Promo2SinceWeek, Promo2SinceYear, PromoInterval`。
34
+ - 随时间变化的 target / 动态协变量 → **measurement**(FIELD)。
35
+ - `timestamp` → `Time`(INT64 毫秒)。
36
+ - 表名:rossmann_1D, rossmann_1W。
37
+
38
+ ### 列含义
39
+
40
+ | 列 | 角色 | TsFile 类型 |
41
+ |---|---|---|
42
+ | `Time` | Time(时间列) | INT64 |
43
+ | `id` | TAG(device 维度) | STRING |
44
+ | `Store` | TAG(device 维度) | DOUBLE |
45
+ | `StoreType` | TAG(device 维度) | STRING |
46
+ | `Assortment` | TAG(device 维度) | STRING |
47
+ | `CompetitionDistance` | TAG(device 维度) | DOUBLE |
48
+ | `CompetitionOpenSinceMonth` | TAG(device 维度) | DOUBLE |
49
+ | `CompetitionOpenSinceYear` | TAG(device 维度) | DOUBLE |
50
+ | `Promo2` | TAG(device 维度) | DOUBLE |
51
+ | `Promo2SinceWeek` | TAG(device 维度) | DOUBLE |
52
+ | `Promo2SinceYear` | TAG(device 维度) | DOUBLE |
53
+ | `PromoInterval` | TAG(device 维度) | STRING |
54
+ | `DayOfWeek` | FIELD(measurement) | FLOAT |
55
+ | `Sales` | FIELD(measurement) | FLOAT |
56
+ | `Customers` | FIELD(measurement) | FLOAT |
57
+ | `Open` | FIELD(measurement) | FLOAT |
58
+ | `Promo` | FIELD(measurement) | FLOAT |
59
+ | `StateHoliday` | FIELD(measurement) | STRING |
60
+ | `SchoolHoliday` | FIELD(measurement) | FLOAT |
61
+
62
+ > 注:有 2230 个原始 id 含非法标识符字符,已规范化为合法 device 名(如 1→_1, 2→_2, 3→_3)。
63
+
64
+ ## 转换说明
65
+
66
+ - 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。
67
+ - 数值类型按源列自适应:float32→FLOAT、float64→DOUBLE、整数→INT64、bool→BOOLEAN。
68
+ - 时间精度:毫秒(INT64)。
69
+ - 大表会被工具自动分片为 `<名>_1.tsfile`、`<名>_2.tsfile` …,同属一个逻辑表。
70
+
71
+ ## 读取示例
72
+
73
+ ```python
74
+ from tsfile import TsFileReader
75
+
76
+ reader = TsFileReader("1D/1D.tsfile")
77
+ schemas = reader.get_all_table_schemas()
78
+ # 表名:rossmann_1D;列见下方"列含义"。
79
+ ```