SalesForecastingWOMartStore (TsFile)

This dataset is the Apache TsFile conversion of SURESHBEEKHANI/SalesForecastingWOMartStore. It preserves the source's retail store and region attributes and its separate training and forecasting tables.

Modalities: Time-series.

Overview

• Source revision: 9fad683234cd403a8e67482eb8250e9c4957f28c
• Training source: TRAIN.csv, 188,340 rows, 365 stores, 2018-01-01 to 2019-05-31 (516 dates)
• Test source: TEST_FINAL.csv, 22,265 rows, 365 stores, 2019-06-01 to 2019-07-31 (61 dates)

TsFile files and schema

Each source split is kept as its own table and TsFile:

File	Table	Rows	Size (bytes)
salesforecastingwomartstore_train.tsfile	salesforecastingwomartstore_train	188,340	2,802,153
salesforecastingwomartstore_test.tsfile	salesforecastingwomartstore_test	22,265	351,910

Both tables use the following key columns:

Column	Role	Type	Source / meaning
Time	TIME	INT64 (ms)	UTC-midnight epoch milliseconds from Date
store_id	TAG	STRING	Source Store_id
store_type	TAG	STRING	Source Store_Type
location_type	TAG	STRING	Source Location_Type
region_code	TAG	STRING	Source Region_Code
id	FIELD	STRING	Source ID
holiday	FIELD	INT64	Source holiday indicator
discount	FIELD	STRING	Source Discount (Yes/No)

The train table additionally contains order_count (INT64, renamed from #Order) and sales (DOUBLE, renamed from Sales). The test table has no target columns because TEST_FINAL.csv does not provide them.

Conversion notes

• Date is parsed to Time at UTC midnight. The redundant source date string is not duplicated as a FIELD.
• Store and region identity columns are string TAGs, so a store/date sequence remains isolated by its complete device key.
• #Order and Sales are renamed to the safe field names order_count and sales only in the train table. No order or sales values are fabricated for test rows.
• All source rows and fields are retained and sorted by TAG columns and Time; duplicate store/date keys are rejected.

Read example

from tsfile import TsFileReader

path = "salesforecastingwomartstore_train.tsfile"
with TsFileReader(path) as reader:
    with reader.query_table(
        "salesforecastingwomartstore_train",
        ["order_count", "sales"],
        batch_size=4096,
    ) as result:
        batch = result.read_arrow_batch()
        if batch is not None:
            print(batch.to_pandas().head())

Source & license

• Original dataset: https://huggingface.co/datasets/SURESHBEEKHANI/SalesForecastingWOMartStore
• Author / publisher: SURESHBEEKHANI
• License: Apache-2.0

Usage

Install the Apache TsFile Python SDK (pip install tsfile) and read a converted file:

from pathlib import Path
from tsfile import TsFileReader

path = Path("salesforecastingwomartstore_test.tsfile")
with TsFileReader(str(path)) as reader:
    schemas = reader.get_all_table_schemas()
    print("tables:", list(schemas))
    table_name = next(iter(schemas))
    table = schemas[table_name]
    columns = [column.get_column_name() for column in table.get_columns()]
    print("columns:", columns)
    field_names = [
        column.get_column_name()
        for column in table.get_columns()
        if column.get_column_name() not in {"Time", "time"}
    ]
    if field_names:
        with reader.query_table(table_name, field_names[:3], batch_size=1024) as result:
            batch = result.read_arrow_batch()
            if batch is not None:
                print(batch.to_pandas().head())