This project is read-only.

ContentItemRecord limitation ?

Topics: Customizing Orchard, Writing modules
Jul 14, 2013 at 5:07 PM
dear all,
i want to develop an application for my company, using orchard.
but there will be a lot of records in this apps.
Since every content creates a record in ContentItemRecord, I wonder if there's potential issue when the data gets bigger and bigger?
or there's any other way to deal with large data in orchard?

Jul 14, 2013 at 6:41 PM
There will be lots of records even on the table you handle, which means if it's not an issue for your tables, it won't for this one.

This extraneous table can add some overhead though, so the only answer that can be given is that you try to add some load to it and see if it complies with your requirements. Please remember you should be able to optimize the db though by using custom indexes where it makes sense.
Jul 16, 2013 at 8:28 AM
yes, you're right.
how do i change the default datatype int for ID to bigint ?
thx before,

Jul 17, 2013 at 12:35 AM
It's a little early to do that. First you must determine that you need it.
Jul 17, 2013 at 2:00 AM
my requirement is that the application should be able to handle 10millions transactions a year.
It's intranet application, so will be used intensively.

do you think orchard is suitable for that? i'm amazed with orchard, so if possible, i would like my apps to be built on it :)

Jul 17, 2013 at 4:29 AM
Why do you need those records to be content items?
Jul 17, 2013 at 6:37 AM
I thought that's how orchard works ?
for example, I would have a content type "Customer", and all of its record would be content items?
If I do it that way, i would be able to extend the type "Customer" later on, such as adding address part, contact person part, etc.

or there's actually another way of doing it other than content item?
I really appreciate your help

Jul 17, 2013 at 8:53 AM
It's the way part of Orchard works, yes, but you can handle records that are not part records. It's hard to say from so little information what the right choice would be in this case. If it's performance that you're worried about, nothing can replace a simulation of the load with a representative dataset.