Indexing multidimensional arrays

Generalised indexing

• Using square brackets to index multidimensional array: X[<Indeces>];
• Indeces: Index[,Index]+;
• Index: {number|array|slice|mask|names};
• Advanced indexing operatos similar to Pandas. E.g. .where, .isin, .query.
• Attaching an index (flat or hierarchical) to a dimention;
• Backends: Numpy, Pandas, H5Py

Numpy

• Indexing with a single element;
• Lists and arrays;
• Slising and striding, negative indexing;
• Structural indexing tools (np.newaxis, ellipsis);
• Broadcasting:

  When operating on two arrays, NumPy compares their shapes element-wise. It starts with the trailing (i.e. rightmost) dimensions and works its way left

• Advanced indexing;
• Masked arrays;
• Array manipulation routines;

References

1. Basic Numpy indexing

Pandas

• Using standard brackets [...];
• Dot (.) operator (.loc, .iloc, .at, .iat);
• Direct attribute access by name;
• Reindexing;
• Setting with enlargement;
• Additional methods: .where, .isin, .query;
• Hierarchical indexing with levels;
• Multi-index cross casting: .form_product, .from_frame, .from_tuples etc.;
• Timeseries support.

References

1. Indexing and selecting data

RLang

• Differentiation between vectors, matrices, lists and data frames;
• Dollar sign operator - $;
• Slicing and striding support, masks, negative indexing.
• Extended operators: which, %in%, match.

References

1. Indexing in R

PyTorch

• Similar to Numpy

References

1. Tensor indexing API