---
license: other
task_categories:
- time-series-forecasting
task_ids:
- univariate-time-series-forecasting
- multivariate-time-series-forecasting
annotations_creators:
- no-annotation
source_datasets:
- original
tags:
- time-series
- forecasting
- benchmark
- fev
- tsfile
- apache-tsfile
- modality:timeseries
- Time-series
- format:tsfile
- arxiv:2509.26468
size_categories:
- 100K<n<1M
pretty_name: walmart (TsFile format)
configs:
- config_name: default
  data_files:
  - split: train
    path: "**/*.tsfile"
---

# walmart (TsFile format)

This repository contains time-series forecasting data stored in [Apache TsFile](https://tsfile.apache.org/) format.

## Summary

- FEV subset: `walmart`
- Unified source collection: [`autogluon/fev_datasets`](https://huggingface.co/datasets/autogluon/fev_datasets)
- Original source: https://www.kaggle.com/competitions/walmart-recruiting-store-sales-forecasting
- Paper / citation: [[24]](https://www.kaggle.com/competitions/walmart-recruiting-store-sales-forecasting/overview/citation)
- Series: 2,936
- Modalities: Time-series
- TsFile rows (flattened observations): 4,609,143
- Frequencies: source-defined
- TsFile files: 1
- Time precision: milliseconds (`INT64`).

Licensing and citation requirements follow the original source. This repository does not claim ownership of the original data.

## Dataset Statistics

| Series | Median series length | TsFile rows (observations) | Dynamic columns | Static columns | Data files |
|---:|---:|---:|---:|---:|---|
| 2,936 | 143 | 4,609,143 | 11 | 4 | `walmart.tsfile` |

## Files

The Hugging Face dataset card YAML points `configs.data_files` to all `*.tsfile` files in this repository.

- `walmart.tsfile`

## TsFile Storage Model

- Each original series (`id`) is stored as one TsFile device.
- Static covariate columns are stored as TAG columns: `Store, Dept, Type, Size`.
- Time-varying targets and dynamic covariates are stored as FIELD measurements.
- Source `timestamp` values are mapped to the TsFile `Time` column as millisecond timestamps.
- Table name(s): walmart.

### Column Schema

| Column | Role | TsFile type |
|---|---|---|
| `Time` | Time column | INT64 |
| `id` | TAG (device dimension) | STRING |
| `Store` | TAG (device dimension) | DOUBLE |
| `Dept` | TAG (device dimension) | DOUBLE |
| `Type` | TAG (device dimension) | STRING |
| `Size` | TAG (device dimension) | DOUBLE |
| `target` | FIELD (measurement) | FLOAT |
| `IsHoliday` | FIELD (measurement) | FLOAT |
| `Temperature` | FIELD (measurement) | FLOAT |
| `Fuel_Price` | FIELD (measurement) | FLOAT |
| `MarkDown1` | FIELD (measurement) | FLOAT |
| `MarkDown2` | FIELD (measurement) | FLOAT |
| `MarkDown3` | FIELD (measurement) | FLOAT |
| `MarkDown4` | FIELD (measurement) | FLOAT |
| `MarkDown5` | FIELD (measurement) | FLOAT |
| `CPI` | FIELD (measurement) | FLOAT |
| `Unemployment` | FIELD (measurement) | FLOAT |

> Note: 2936 original `id` values contained invalid identifier characters and were normalized to valid device names, for example 10_1→_10_1, 10_10→_10_10, 10_11→_10_11.

## Conversion Notes

- The source FEV format stores each time series as one nested row containing `id`, `timestamp[]`, and target or covariate arrays.
- The TsFile conversion flattens those nested arrays into long rows. Therefore, the `TsFile rows` values above correspond to the number of timestamped observations after flattening.
- TAG columns identify the device and static metadata. FIELD columns contain values that change over time.
- Large logical tables may be split into multiple `.tsfile` shards such as `<name>_1.tsfile`, `<name>_2.tsfile`, and so on. Shards listed for the same frequency belong to the same logical table.

## Reading Example

```python
from tsfile import TsFileReader

reader = TsFileReader("walmart.tsfile")
schemas = reader.get_all_table_schemas()
# Table name(s): walmart
```