Programming Pig, 2nd Edition

Chapter 4. Introduction to Pig Latin

It is time to dig into Pig Latin. This chapter provides you with the basics of Pig Latin, enough to write your first useful scripts. More advanced features of Pig Latin are covered in Chapter 5.

Preliminary Matters

Pig Latin is a data flow language. Each processing step results in a new dataset, or relation. In input = load 'data', input is the name of the relation that results from loading the dataset data. A relation name is referred to as an alias. Relation names look like variables, but they are not. Once made, an assignment is permanent. It is possible to reuse relation names; for example, this is legitimate:

A = load 'NYSE_dividends' (exchange, symbol, date, dividends);
A = filter A by dividends > 0;
A = foreach A generate UPPER(symbol);

However, it is not recommended. It looks here as if you are reassigning A, but really you are creating new relations called A, and losing track of the old relations called A. Pig is smart enough to keep up, but it still is not a good practice. It leads to confusion when trying to read your programs (which A am I referring to?) and when reading error messages.

In addition to relation names, Pig Latin also has field names. They name a field (or column) in a relation. In the previous snippet of Pig Latin, dividends and symbol are examples of field names. These are somewhat like variables in that they will contain a different value for each record as it passes through the pipeline, but you cannot assign values to ...