Home
Introduction
The 1010data Macro Language is the most powerful and direct way to interact with 1010data's platform.
Query optimization
Often there is more than one solution to a single problem, and depending on the problem and the data, different approaches can produce better performance. It is important to understand what impacts query performance and how to translate that into optimizing your own queries.
Links
You can use <link> operations in 1010data to combine and filter data between tables. However, depending on the type of link and size of your data, this operation can be costly. It is important to know what steps can be taken while using <link> to optimize your query performance.
Prelinks
You can save time when linking two tables where a prelink exists because the <link> is already evaluated.

Introduction
The 1010data Macro Language is the most powerful and direct way to interact with 1010data's platform.
- What is in a recipe?
  Each recipe in this Cookbook, with the exception of the Debugging Guide, contains the same basic elements. This page describes the contents of each section contained in a recipe and the elements that make a Debugging Recipe different.
- Difficulty criteria
  Each recipe is equipped with a difficulty measure based on the criteria discussed in this section. A recipe holding a specific difficulty rating can contain one more of the elements in the description of that level or a lower level.
- Query optimization
  Often there is more than one solution to a single problem, and depending on the problem and the data, different approaches can produce better performance. It is important to understand what impacts query performance and how to translate that into optimizing your own queries.
  - Lazy evaluation
    Certain operations in 1010data are not evaluated until the data from those operations are needed by another part of the query.
  - Selections
    The order in which selections are made and the method in which selections are made can affect the efficiency of your query.
  - Links
    You can use <link> operations in 1010data to combine and filter data between tables. However, depending on the type of link and size of your data, this operation can be costly. It is important to know what steps can be taken while using <link> to optimize your query performance.
    - Tables vs. worksheets
      When you perform one or more operations on a table, it becomes a worksheet. It is important to understand the differences and implications of linking to a worksheet rather than a table.
    - Prelinks
      You can save time when linking two tables where a prelink exists because the <link> is already evaluated.
    - Segby links
      Linking on a column that your foreign table is both sorted and segmented by (i.e., sortseg), can make your <link> more efficient.

Prelinks

You can save time when linking two tables where a prelink exists because the <link> is already evaluated.

Key concepts

Prelinks optimize future query performance.
Prelinks can be created through the data loading process or through the query language.

Discussion

A prelink creates a vector of index values for each row in the base table that corresponds to the matching row in the foreign table.

To create a prelink, you can use materialize="1" within the <link> operation.

Note: You must own the table and have materialize access in order to use materialize="1" to create a prelink.

Prelinks between tables can also be created during the data loading process.

Prelinks are useful when you link the same tables together often. It allows you to evaluate the <link> once, when the prelink is created, and then take advantage of the evaluation whenever you link those tables together.

Example

You often link the Sales Detail table (pub.doc.retail.altseg.sales_detail_sku) to the Product Master table (pub.doc.retail.altseg.products) in order to gain more information on the products. Due to the size of your transaction table, these <link> operations can be expensive. In order to save time on future queries, you create a prelink from the Sales Detail to the Product Master.

<base table="pub.doc.retail.altseg.sales_detail_sku"/>
<link table2="pub.doc.retail.altseg.products" col="sku" col2="sku" materialize="1"/>

Now there exists a file in the base table that holds information on the mapping between the two tables. Next time you link these two tables, the file is read to determine the match for each row instead of searching the foreign table for each match.