Introduction

YQL (Yandex Query Language) is a language of universal declarative queries against data storage and processing systems as well as an infrastructure to run such queries. YQL benefits include:

  • A powerful graph execution engine that can build MapReduce pipelines with hundreds of nodes and adapt during computation.
  • Ability to build complex data processing pipelines using SQL by storing subqueries in variables as chains of dependent queries and transactions.
  • Predictable parallel execution of queries of any complexity.
  • Efficient implementation of joins, subqueries, and window functions with no restrictions on their topology or nesting.
  • Extensive function library.
  • Support for user-defined functions in C++, Python.
  • Automatic execution of small parts of queries on prepared compute instances, bypassing MapReduce operations to reduce latency.

YQL provides a functional web interface where, among other things, you can:

  • Write query code.
  • Start and stop query execution.
  • View query execution results.
  • View query history.

How to try

To run your first YQL query:

  1. Open the web interface of the YTsaurus cluster and go to the Queries tab — you can get there from the menu on the left.

  2. Enter the query and click the start button:

    SELECT "Hello, World!";

Glossary

Term

English translation

Description

Query

Query

Program text in YQL

Operation

Operation

Query execution process. In YQL terminology, a query and an operation relate to each other much like a program and a process do in an operating system.

Table

Table

Logically, a table is a list of structures (List<Struct<...>>) that is a basic primitive for processing large volumes of data in YQL.

List items form rows, while structure members serve as table cells. Vertically aligned cells constitute a table column.

Expression

Expression

A computed value. Typically, it takes one or more table cells as input, with the output becoming a cell in another table. Examples:

  • 2 * 2;
  • String::ToUpper("hi, " \|\| name \|\| "!")

Statement

Statement

Query components separated by semicolons and starting with a verb. Examples:

  • SELECT ...;
  • INSERT INTO ... SELECT ... ;
  • REDUCE ... USING ... ON ...;

Subquery

Subquery

A query component that, similarly to tables, can be used as input for statements or other subqueries. Optionally, subqueries can be parameterized

Named node

Named node

A mechanism for reusing expressions and subqueries within one query. More

Examples:

  • $foo = 2 + 2;
  • $bar = (SELECT bar FROM my_table);

Lambda

Lambda

A parameterizable block consisting of one or more named nodes (specifically expressions, not statements) where the result of the last node becomes the result of the entire lambda call. The call is made by passing parameters in parentheses. Learn more

UDF (User-Defined Function)

UDF (User-Defined Function)

Functions that let you integrate business logic into a query using one of the supported popular programming languages.

C++ UDFs are loaded in compiled .so format, with a wide range of built-in functions available as well. Python UDFs can use a script from any query string, but require additional description of their signature.

Since the YQL optimizer can't peek inside the Python interpreter, we recommend using lambda functions and C++ UDFs for better performance.

Action

Action

A parameterizable block consisting of one or multiple statements that can then be invoked any number of times using special keywords. Unlike lambda functions, they don't return any result. Learn more

Library

Library

A query component stored in a separate file for reuse or convenience. Learn more
Previous
Best practices
Next
Overview