Slow performance while accessing Outlook item fields using VSTO

[maxim]

Hi,

I'm developing a C# based Outlook add-in using VSTO 2005. This add-in
collects information from multiple Mail items, so I've a simple code like
this:

Outlook.MAPIFolder inbox =
this.Application.ActiveExplorer().Session.GetDefaultFolder(Outlook.OlDefaultFolders.olFolderInbox);

Outlook.Items items = inbox.Items;

foreach(Outlook.MailItem mailItem in items) {

resultList.Add(new FoundMessageItem(mailItem)); //Takes some fields from
the mailItem like subject, sender, recipient

}

Running this loop on even quite small mailbox (about 100 items) takes a long
time (about 10-12 seconds) which is unexptable in my case. What can be done
to improve the performance?

Thanks in advance,

Maxim

 

[Ken Slovak - [MVP - Outlook]]

Interop code will always be slower than straight COM code, it's essentially
late bound and has all the overhead of the Interop and Framework. You can
try using SetColumns to reduce the load time for each item in the Items
collection (remember to clear that with ResetColumns when done) setting only
the properties you need to work with.

Other than that you'd need a faster API, such as CDO 1.21 or Extended MAPI,
but neither is supported for .NET code.

FWIW, if you ever run into non-mail items in Inbox your code will throw and
exception and crash. You can have various other types of items there such as
Posts, NDR's, meeting and task requests. Testing for the class of an item
before casting it to a MailItem or enclosing the loop in a try...catch block
would prevent the exception but then of course the code will be even slower.

 

[maxim]

Ken, thank you very much for your fast response.
I changed my code slightly to use CType paradigm of the VB.NET language, so
my code now looks like:
For Idx = 1 To items.Count
item = items.GetNext()
subject = CType(item.Subject, String)
entryId = CType(item.EntryID, String)
Next
This loop works much faster than the C# analog. If I retrieve only Subject
field, the performance is fantastic but it became much slower when I add
additional fields like EntryId. Can you explain me pls, why there is so big
difference between retrieving only Subject field and retrieving Subject and
EntryId fields?
Does Outlook performs requests to the Exchange server for retrieving fields
other than Subject? I'm working with Outlook2003.
Thanks,
Maxim

 

[Ken Slovak - [MVP - Outlook]]

Retrieving 2 fields should be 1/2 as fast as retrieving 1 <g>

When you get an Item (in any way) Outlook loads all the properties for that
item. There are special cases like cached mode download only headers and so
on, but in general that's the behavior.

If you use SetColumns to only load the properties you want things are much
faster. Instead of loading everything the underlying MAPI is only loading
those columns into each row of the Items table that you requested. That
provides a large performance boost. Of course properties that are not in the
SetColumns call are not available.

I've never made any speed comparisons as to what type of loop is faster in
various .NET languages. I do know that in VBA/VB6 for example that
For...Each is faster than in indexed For loop. So different loop constructs
can vary greatly in speed depending on their implementation. For all I know
a do loop would be even faster (or it might be slower). Unless you can find
the metrics for different types of loops you'd need to discover that for
yourself. The C# groups might have answers for that sort of metrics.