Try our conversational search powered by Generative AI!

Daniel Ovaska
Mar 5, 2019
  3548
(2 votes)

Performance - GetChildren vs GetBySegment

GetChildren is a decent method that is also cached in the background. But if you have 1000s of children you will get some performance hits. Another good option if you are only searching for a single content item is to use GetBySegment method. This one works well with large amounts of children with excellent performance. 

I tried it on folder that has 13000+ children (yes, bad idea) and these were the results when running on local machine

  1. GetChildren
    
    foreach (var folder in ContentRepository.GetChildren<ContentFolder>(parent)​)
    {
       if (folder.Name == customerId)
       {
          ...                    
       }
    }
    17.857s
  2. GetBySegment
     var customerFolder = ContentRepository.GetBySegment(parent, customerId, LanguageSelector.AutoDetect());​
    0.147s

So that's a pretty huge performance gain by more than a factor 100. Hope it helps someone out there to pick the right tool for the job. 

Few children => GetChildren
Many children => GetBySegment or Episerver Find...

Happy coding everyone!

Mar 05, 2019

Comments

K Khan
K Khan Mar 5, 2019 02:29 PM

Thanks for sharing, any suggestion for contentRepository.GetAncestors?

Praful Jangid
Praful Jangid Mar 5, 2019 03:07 PM

So, is this GetBySegment(...) works as a filter by ContentId you passed on (second parameter)?

But, if I see the definition here, that is accepting urlSegment as second parameter. What is that? Or do we have overloaded functions for this?

Daniel Ovaska
Daniel Ovaska Mar 5, 2019 03:17 PM

K Khan: Don't think that method will be a problem. Unless you have very very deep tree structure. Have you seen any performance issues with that one?

Praful Jangid: It's a filter but on the url segment yes. Url segment is the part of the url for that page that is visible in browser. For instance, the url segment for this blog post is performance---getchildren as you can see above in the browser. So if you know the url segment (or route segment as it's also called) you can use that to fetch the page.

Daniel Ovaska
Daniel Ovaska Mar 5, 2019 03:22 PM

So for instance, if I'm building this blog on Episerver world and I want get the 2019 folder for blogs and I have this structure

https://world.episerver.com/blogs/Daniel-Ovaska/Dates/2019/3/performance---getchildren/

If I have the parent page with url 

https://world.episerver.com/blogs/Daniel-Ovaska/Dates/

I can then use contentRepo.GetBySegment(parent,"2019",LanguageManager.AutoDetect()) to get the page with url

https://world.episerver.com/blogs/Daniel-Ovaska/Dates/2019/

Erik Norberg
Erik Norberg Mar 6, 2019 01:37 PM

If you are going to use GetBysegment you have to use caution.

I recently filed a bug about it that has been deemed "by design". If GetBySegment doesn't get a match in the provided language it will return a match for any other language it can find. It might not always be the expected or desired behaviour.

Daniel Ovaska
Daniel Ovaska Mar 14, 2019 08:52 AM

Good to know! 

Please login to comment.
Latest blogs
New Series: Building a .NET Core headless site on Optimizely Graph and SaaS CMS

Welcome to this new multi-post series where you can follow along as I indulge in yet another crazy experiment: Can we make our beloved Alloy site r...

Allan Thraen | Jun 14, 2024 | Syndicated blog

Inspect In Index is finally back

EPiCode.InspectInIndex was released 9 years ago . The Search and Navigation addon is now finally upgraded to support Optimizely CMS 12....

Haakon Peder Haugsten | Jun 14, 2024

Change the IP HTTP Header used for geo-lookup in Application Insights

.

Johan Kronberg | Jun 10, 2024 | Syndicated blog

Copying property values

In this article I’d like to show simple Edit Mode extension for copying property values to other language versions. In one of my previous blogposts...

Grzegorz Wiecheć | Jun 8, 2024 | Syndicated blog