pydata/xarray
Xarray
Xarray 是一个 Python 多维数组库和带标签的数据集框架。它扩展了 NumPy 数据结构，通过为数组添加标签，允许使用命名维度和坐标来组织复杂的 N 维数据。
该库提供了 NetCDF 数据接口，用于读写 NetCDF 和 Zarr 等科学数据格式。它通过在数学运算过程中维护数据与物理坐标之间的关系，实现了科学数组计算。
该项目涵盖了多维数据分析、地理空间数据处理和气候数据处理。它还支持用于管理随时间变化的观测序列的时间序列分析。
Features
- Labeled Array Management - Provides a framework for organizing multidimensional datasets using named dimensions and coordinates to avoid integer indexing.
- Labeled Dataset Aggregations - Groups multiple labeled arrays into a single container using a shared set of coordinates and global attributes.
- Labeled Dataset Frameworks - Provides a system for organizing complex multidimensional data with metadata and coordinate-based alignment.
- Dimension Coordinate Mapping - Maps named dimensions to integer axis indices to automate coordinate alignment and array slicing.
- Multidimensional Arrays - Enables analysis of large arrays of data using named dimensions and labels instead of raw integer indexing.
- Scientific Computing - Performs mathematical operations on labeled N-dimensional arrays while maintaining the relationship between data and physical coordinates.
- N-Dimensional Array Libraries - Provides a library for managing labeled N-dimensional arrays and datasets using named dimensions and coordinates.
- NumPy Array Integration - Uses NumPy arrays as the underlying data container to leverage efficient contiguous memory and vectorised operations.
- Labeled Array Extensions - Extends the NumPy data structure by adding labels to arrays to simplify indexing and data analysis.
- Deferred Computation Integration - Wraps Dask arrays to defer computation until a result is explicitly requested.
- Time Series Analysis - Manages sequences of multidimensional observations over time, allowing for easy resampling and alignment of temporal data.
- Array Broadcasting - Implements array broadcasting to automatically align datasets by matching labels across dimensions during arithmetic operations.
- Climate Data Processing - Handles complex atmospheric or oceanic datasets that use coordinates like latitude, longitude, and time.
- Geospatial Data Manipulation - Works with spatially referenced grids and datasets to slice, dice, and align data based on geographic coordinates.
- Scientific Data Export Formats - Provides a data interface for reading and writing scientific data formats such as NetCDF and Zarr.
- Data Containers - N-dimensional labeled arrays and datasets for complex data.
- Data Manipulation - Multidimensional data selection and analysis.
- Data Processing Libraries - N-dimensional array structures with labeled dimensions.
