It now includes Node type, binaryvar, binaryudt, spatial, timeseries, Websphere MQ.
World is full of hierarchies. You have a boss, so does your boss. In pharmaceutical industry, they establish hierarchical relationship for boxes, containers and pallets of medicine. If you want to recall a specific medicine, it's easier to find all the containers and destinations. Given an item and container, It also becomes easier to find other items in the same container. Node datatype helps you to model and query these hierarchical relationships. Check out the working examples in chapter 9 of this redbook, a article from the original designers and of course, Informix Manuals. With 11.50.xC5, Informix also supports CONNECT BY syntax to discover hierarchical relationships within relational table without using node datatype.
If you think BYTE and BLOB are too large for your binary data, binaryvar and binaryudt will help you deal with smaller data. Checkout examples at our redbook. These types provide new types and new functions to operation on these types. Check out the working examples in chapter 9 of this redbook.
Spatial datatype helps you to model spatial data, duh. With Informix, you also get spatial index based on R-tree. Apart from seamless ESRI integration, Informix implementation has two things going for it. Inventor of R-Tree, Toni Guttman worked for Informix and help implement this and we've added statistics collection and costing model to choose the cost effective R-TREE index or other available access path to the table. Tidbit: Informix 11.70 has multi-index scan. Multi-index scan access method is implemented for B-tree indices and not considered for R-tree.
Time series saves you space while dramatically improving performance of data storage and analysis when collect large sensor or other time variant data set. See Informix chief architect, Kevin brown's blog on this use case.
...One more thing...
The power of open C-Lucene text search is available on Informix as Basic text search index. Again, checkout the complete examples in the chapter 9 of the redbook.
In Informix v11.70, you simply do:
CREATE TABLE article_tab(id integer, title lvarchar(512));
CREATE INDEX title_index ON article_tab(title bts_lvarchar_ops) USING bts;
-- only two changes from a simple B-tree index creation. operator class for the column and then USING class as bts, which stands for Basic Text Search
-- load the data
-- start searching!
select id from article_tab where bts_contains(title, 'informix');
(see full example in the redbook)
Enjoy the new types!
If you've used all these in Informix for a while and wondering what's new, here's the answer.
First, thank you for learning and using all the datablade registration, maintenance, etc.
Informix v11.70 still supports datablade registration/etc, if you prefer to use it.
However, all these new types, access methods are known to Informix.
On first use of new types or access methods, Informix v11.70, silently loads these datatypes, creates any prerequisites.
So, using these data types, indices became whole lot easier.