Data Tables
Greg Wilson · 2022
Modern software systems are so complex that it&s;s practically impossible to predict their performance in advance. Instead, developers need to conduct some experiments to find out what&s;s fast or slow in practice. To see how to do this systematically, this chapter looks at several ways to implement data tables with one or more named columns and zero or more rows. Tables like appear over and over again in programming, from spreadsheets and databases to the dataframes of R and Python. The key operations on data tables are those provided by SQL: filter, select, summarize, and join. These can be implemented in about five hundred lines of code, but their performance varies dramatically depending on how the data is stored.