Tous les livres

Professionnel

Applis À propos Coach Connexion Commencer la lecture

Quantitative Finance · Glossaire

Qu'est-ce que « Row group, column chunk, data page » ?

Aussi appelé : row group · column chunk · data page

Definition 3.4 Research, Data and Risk Platforms · Chapitre 3 — Columnar Formats

A row group is a horizontal slice of a columnar file’s rows, stored as one column chunk per column, each contiguous in the file. A column chunk is divided into data pages, the unit of encoding and compression. The file’s metadata (its schema, and the position, size, encoding and statistics of every chunk) is written in a footer after the data, so that the file can be written in one pass and read from its end.

Anatomy of a file in the on-disk columnar format: row groups of column chunks, each chunk a sequence of data pages, and a footer after the data that says where every chunk is and what range of values it holds.
Figure 3.3. Anatomy of a file in the on-disk columnar format: row groups of column chunks, each chunk a sequence of data pages, and a footer after the data that says where every chunk is and what range of values it holds.
Lire dans le chapitre →