next »

[konstantin]

I have several category of news and each category have thousands news.

Can Ditto help me?

[Andy Ayre]

I think you might run into performance problems with that many documents. Probably a better approach would be to create a custom MySQL table, import your articles and then modify Ditto or write your own snippet to display them. Shouldn't be hard.

Andy

[Mark]

Ditto 1.1b1 or later will work fine with this. Hopefully, by 1.1 performance will be even better.

You might consider paging the results to make it easier on the visitor--- and your db server.

[Andy Ayre]

Ditto will work fine with 1000s of documents or will work fine with a custom table?
I thought that MODx itself had performance problems above 5000 documents, so I was referring to that rather than any performance problems with Ditto.

Andy

[Mark]

Ditto will work with as many documents as MODx supports. So, if he has less than 5k he will be fine. If he has more than that, it is possible to modify Ditto to use an alternate table but it would require considerable changes to 2 core functions.

[OpenGeek]

Ditto will work fine with 1000s of documents or will work fine with a custom table?
I thought that MODx itself had performance problems above 5000 documents, so I was referring to that rather than any performance problems with Ditto.

Andy, this has nothing specifically to do with Ditto actually, though I'm sure Ditto adds some additional overhead to the problem.  The real problem is the current approach to caching the entire site structure within a single cache file in MODx.  Sites with thousands of documents will have very large siteCache files which will tax the web server more by requiring more memory, more time to load the file from the filesystem, and more time to process it with the PHP interpreter.  In addition, thousands of pageCache files will also exist in a single directory, which can also increase resource loads for the web server when reading the cache directory contents searching for a cached file.  This reduces performance exponentially.

All of this is being addressed at the core level, and hopefully we'll have some additional scalability in this respect with the next minor release of MODx.   

[Andy Ayre]

Yes, I understand, that's why I mentioned I was referring to MODx as the limitation, not Ditto. There isn't much point trying to use Ditto with 5k+ documents if MODx can't handle it well.

Andy

[mooreds]

Ditto will work fine with 1000s of documents or will work fine with a custom table?
I thought that MODx itself had performance problems above 5000 documents, so I was referring to that rather than any performance problems with Ditto.

Andy, this has nothing specifically to do with Ditto actually, though I'm sure Ditto adds some additional overhead to the problem.  The real problem is the current approach to caching the entire site structure within a single cache file in MODx.  Sites with thousands of documents will have very large siteCache files which will tax the web server more by requiring more memory, more time to load the file from the filesystem, and more time to process it with the PHP interpreter.  In addition, thousands of pageCache files will also exist in a single directory, which can also increase resource loads for the web server when reading the cache directory contents searching for a cached file.  This reduces performance exponentially.

All of this is being addressed at the core level, and hopefully we'll have some additional scalability in this respect with the next minor release of MODx.   

I'm curious--can you get around the 5K limit on total documents via turning off caching?

Thanks,
Dan

[OpenGeek]

All of this is being addressed at the core level, and hopefully we'll have some additional scalability in this respect with the next minor release of MODx.   

I'm curious--can you get around the 5K limit on total documents via turning off caching?

Thanks,
Dan

Not yet Dan; part of the bottleneck is a large index of document metadata that is loaded on all requests, which a large part of the current parsing engine is dependent on.  As I mentioned though, these issues are being addressed.