When using FetchXML query to read data from Microsoft Dynamics CRM server, you might be following the SDK document to page through the query results in order to get all records that satisfy the query. You might run into a situation that even you have supplied the paging cookie by following the documentation, you are still receiving an error complaining "Paging cookie is required when trying to retrieve a set of records on any high pages".
You might be pulling your hair when this happens, since you would think you had done everything right, and you have followed the exact code sample provided in the SDK sample.
If you are wondering why, you are actually not alone. I have run into this error more than a couple of times while supporting our clients who use our CRM source component. With our CRM Source Component, we do page through the entire recordset based on the query provided, and we always use page-cookies in subsequent queries, so there should be no reason for this error to happen. I was pulling my hair too when I first saw this error. But I did spot something while looking at our client's query, which leads to a solution for the problem, so I thought it worth writing a blog post which could save your some research.
If you ever run into this, the first thing that you should check is to make sure that you have included the primary key field of the primary entity in the query, second thing that you want to check whether you have used a distinct operator in your query. Say that you have a query like the following, you would want to ask yourself if you really intend to do a distinct on a large number of CRM records (say anything more than the page size that you are using - we call it the Batch Size in our CRM Source Component).
<fetch version="1.0" output-format="xml-platform" mapping="logical" distinct="true"> <entity name="account"> <attribute name="name" /> <attribute name="address1_line1" /> <attribute name="address1_line2" /> <attribute name="address1_line3" /> <attribute name="address1_city" /> <attribute name="address1_stateorprovince" /> <attribute name="address1_postalcode" /> <attribute name="telephone1" /> <order attribute="name" descending="false" /> </entity> </fetch>
When the distinct operator is not used in a FetchXML query, CRM server would typically automatically add the primary key field to the query and return back the IDs, which will be used as part of the paging cookie for the next service call. But if the query contains a distinct operator, this will not happen, which is the reason that we are getting this error message.
With that in mind, the solution for the error is rather simple. We have a couple of options to fix the issue.
- You can add the primary key field to the query. As your original intention is to do a distinct, you would want to ask yourself if adding the primary key field invalidates your original intention.
- You can remove the distinct operator from your FetchXML query. Again, this is a different change compared to the above, you would want to ask yourself whether the new query would satisfy your requirements.
If it happens that you don't like either of the above options, you will have to program your paging code, so you check if MoreRecords is true and also the returned PagingCookie is not empty as the combined conditions to decide whether to navigate to the next page.
Hope this helps.