Home
Extending Informix®
Beyond standard relational database objects, HCL Informix® can be extended to handle specialized data types, access methods, routines, and other objects. Informix includes many built-in extensions that are fully integrated in the database server. Informix also provides DataBlade® modules, which are packages of extended database objects for a particular purpose and that are installed separately from the database server. Alternatively, you can create your own user-defined objects for Informix.
Creating extensions
You can create user-defined data types, routines, access methods, and other database objects to suit your needs. You can use application programming interfaces to write user-defined routines and applications that access data in Informix® databases.
DataBlade® API Programmer's Guide
The Informix® DataBlade® API Programmer's Guide describes information about the DataBlade API, the C-language application programming interface (API) provided with HCL Informix.
Create user-defined routines
Extend data types
This section describes the ways to extend data types with C user-defined routines (UDRs).
Provide statistics data for a column
The database server can provide statistics data for the columns of a table. This statistics data describes the distribution of the values within a column. The query optimizer uses this statistics data to determine the best path for an SQL statement. With this information, the optimizer can estimate the effect of a WHERE clause by examining, for each column included in the WHERE clause, the proportionate occurrence of data values contained in the column.
Collect statistics data
The UPDATE STATISTICS statement collects statistics about the tables in your database. It automatically collects statistics for all columns with built-in data types (except TEXT and BYTE). However, it cannot automatically collect statistics for columns with user-defined data types because it does not know the structure of these data types.
Design the user-defined statistics
Before you begin to code a statcollect() function for a particular user-defined data type, you need to decide what it means to collect statistics on this data type.

Extending Informix®
Beyond standard relational database objects, HCL Informix® can be extended to handle specialized data types, access methods, routines, and other objects. Informix includes many built-in extensions that are fully integrated in the database server. Informix also provides DataBlade® modules, which are packages of extended database objects for a particular purpose and that are installed separately from the database server. Alternatively, you can create your own user-defined objects for Informix.
- Informix® extensions and DataBlade® modules
  These topics describe how to use built-in database extensions and separately installed DataBlade® modules.
- Creating extensions
  You can create user-defined data types, routines, access methods, and other database objects to suit your needs. You can use application programming interfaces to write user-defined routines and applications that access data in Informix® databases.
  - DataBlade® API Programmer's Guide
    The Informix® DataBlade® API Programmer's Guide describes information about the DataBlade API, the C-language application programming interface (API) provided with HCL Informix.
    - DataBlade® API overview
    - Data manipulation
    - Database access
    - Create user-defined routines
      - Develop a user-defined routine
        A C user-defined routine (UDR) is a UDR that is written in the C language and uses the server-side implementation of the DataBlade® API to communicate with the database server. C UDRs (functions and procedures) are implemented as C-language functions. DataBlade modules often include C UDRs that are made available for use by registering them in the database.
      - Write a user-defined routine
        This section outlines some implementation issues for C user-defined routines (UDRs). This section also describes the specific features and tasks of a C UDR.
      - Manage memory
        A C user-defined routine (UDR) has access to shared memory for dynamic allocations and stack memory for routine arguments (including the MI_FPARAM structure), local stack variables, and return values.
      - Create special-purpose UDRs
        This section describes how to create special UDRs, such as iterator functions, user-defined aggregates, and optimization functions.
      - Extend data types
        This section describes the ways to extend data types with C user-defined routines (UDRs).
        Create an Opaque Data Type
        This section describes how to design and write an opaque data type.
        Provide statistics data for a column
        The database server can provide statistics data for the columns of a table. This statistics data describes the distribution of the values within a column. The query optimizer uses this statistics data to determine the best path for an SQL statement. With this information, the optimizer can estimate the effect of a WHERE clause by examining, for each column included in the WHERE clause, the proportionate occurrence of data values contained in the column.
        Collect statistics data
        The UPDATE STATISTICS statement collects statistics about the tables in your database. It automatically collects statistics for all columns with built-in data types (except TEXT and BYTE). However, it cannot automatically collect statistics for columns with user-defined data types because it does not know the structure of these data types.
        Design the user-defined statistics
        Before you begin to code a statcollect() function for a particular user-defined data type, you need to decide what it means to collect statistics on this data type.
        Define the statistics-collection function
        Collect the statistics
        The statcollect() user-defined function is an iterator function; that is, the database server calls statcollect() for each of the rows on whose column of a user-defined data type UPDATE STATISTICS is collecting statistics. As with other iterator functions, the database server uses an iterator-status constant to indicate when the statistics-collection function is called.
        Register the statcollect() function
        As with any user-defined function, you register the statistics-collection function with the CREATE FUNCTION statement.
        Execute the UPDATE STATISTICS statement
        To collect user-defined statistics, run the UPDATE STATISTICS statement in HIGH or MEDIUM mode.
        User-defined statistics
        Optimize queries
    - Appendix
      This section contains additional reference information.
  - DataBlade® API Function Reference
    The Informix® DataBlade® API Function Reference describes the DataBlade API functions and the subset of Informix ESQL/C functions that the DataBlade API supports.
  - DataBlade® Developers Kit
    The DataBlade® Developers Kit (DBDK) guides contain information about the tools you can use to develop and package DataBlade modules, which extend the functionality of HCL Informix® databases.
  - J/Foundation Developer's Guide
    The J/Foundation Developer's Guide describes how to write user-defined routines (UDRs) in the Java™ programming language for Informix®.
  - R-Tree Index User's Guide
    The Informix® R-Tree Index User's Guide describes the HCL Informix R-tree secondary access method and how to access and use its components.
  - User-Defined Routines and Data Types Developer's Guide
    The Informix® User-Defined Routines and Data Types Developer's Guide describes how to define new data types and enable user-defined routines (UDRs) to extend HCL Informix.
  - Virtual-Table Interface Guide
    The Informix® Virtual-Table Interface Programmer's Guide explains how to create a primary access method with the Virtual-Table Interface (VTI) so that users have a single SQL interface to HCL Informix tables and to data that does not conform to the storage scheme of HCL Informix.
  - Virtual-Index Interface Guide
    The Informix® Virtual-Index Interface Programmer's Guide explains how to create a secondary access method with the Virtual-Index Interface (VII) to extend the built-in indexing schemes of HCL Informix typically with a DataBlade® module.

Design the user-defined statistics

Before you begin to code a statcollect() function for a particular user-defined data type, you need to decide what it means to collect statistics on this data type.

For example, consider the following issues:

Do the values of the user-defined type have some ordering?
To be able to group the values into bins of related values, the data must have some kind of implied sequence. A common use of statistics information is within a selectivity function for a query filter such as “less than" or “greater than". If the values of the user-defined data type do not have ordering, they would not logically be used in such filters. For more info, see Query selectivity.
How does the distribution handle SQL NULL values?
For example, the distribution can ignore NULL values or it can aggregate them. However, the handling of the NULL values must make sense to the user-defined data type.