ArcherYjl/data_sample_1000

TAAC2026演示数据集包含1000个用户-物品交互记录，用于TAAC2026竞赛。数据集采用扁平列布局，所有特征作为顶级列存储，而非嵌套结构。总共有120列，分为6个类别：ID与标签（5列，包括用户ID、物品ID、标签类型、标签时间和时间戳，无空值）、用户整数特征（46列，包括标量和数组类型）、用户密集特征（10列，为浮点数组）、物品整数特征（14列，包括标量和数组类型）、领域序列特征（45列，来自4个行为领域的整数序列）。文件为demo_1000.parquet，大小约39 MB。

The TAAC2026 demo dataset contains 1000 user-item interaction records for the TAAC2026 competition. The dataset uses a flat column layout, with all features stored as individual top-level columns instead of nested structs/arrays. There are a total of 120 columns, categorized into 6 groups: ID & Label (5 columns, including user_id, item_id, label_type, label_time, and timestamp, with no null values), User Int Features (46 columns, scalar or array types), User Dense Features (10 columns, float arrays), Item Int Features (14 columns, scalar or array types), and Domain Sequence Features (45 columns, integer sequences from 4 behavioral domains). The file is demo_1000.parquet, with a size of approximately 39 MB.