Large Numbers of Categories

[Mach325]

Hi,

Before I start writing lots of code to import my category structure, please can you let me know if Imatch 5 would have issues with more than 300,000 categories (UK only)? This number is likely to increase substantially when I add the USA and France.

Thank you.

Neil

[Mario]

I test IMatch IMatch with up to 50,000 categories, which is more than sufficient even for users with thousands of keywords, regular categories, multiple multi-level data-driven categories.

300,000 categories and more is out of specification. Why do you need so many categories?
All this information not only has to be stored, bot loaded, kept cached in memory, traversed, updated etc.
And I doubt that Microsoft has constructed the Windows tree control element to handle that massive amount of data. I already perform a lot of tricks to keep it snappy for 50,000 nodes.

[Mach325]

Hi

There are over 40,000 uniquely named cities, towns and villages in the UK; there are at least 218,308 uniquely named residential streets in the UK (with more being added every month); there are thousands upon thousands of uniquely named buildings/bridges/structures; there are thousands upon thousands of historic sites. It doesn't take much to top 300,000 unique categories/keywords.

If 50,000 is the limit for IMATCH, I will ensure that I don't go above this figure.

Thank you.

[Mario]

It may be that there are so many cities, buildings or post boxes. But do you have images for each of those?

There is no fixed limit. But I only test up to 50,000. I once had 80,000 because a data-driven category was accidentally created with date/time including seconds and that gave a lot of extra categories. It worked and performance was not affected. But I don't test it regularly. If you have a schema for that amount of data, please send it to me so I can maybe run a test. I think that the Windows tree control will break first. But IMatch only loads the categories you really unfold so this may also be able to handle that many entries. IMatch keeps categories in memory for performance, but the memory usage is moderate.