What is the difference between a label and and the field itself? If it’s a matter of too many labels, it still doesn’t really help me understand why the entire RDD needed to be recreated, which includes more labels that my original custom RDD did.
The data for a field is extracted from the Transactional DB and is inserted into the Reporting DB in a new table created based on the Schema defined in the RDD. For each table defined in the RDD a table is created in the Reporting DB and while Epicor tables can be fairly wide, none of them are approaching the 1024 column limit.
The Epicor standard is not to include the text for any field in the SSRS report definition itself. Instead, the text output on a report is extracted from the Report Label Table which is created during the report data extraction phase. The Report Label Table contains a column for every field defined in the report where the label is not excluded and in reports with numerous tables, hitting the limit is a possibility. Epicor uses this method as a way to have one report definition regardless of the language in which the report will be printed. The Report Label table includes a row for every installed language and the RDL will select the Label row for the language “in scope”.
During Upgrade processing, Custom RDDs based on Epicor RDDs get “fixed up” with the addition of any new Tables / Columns / Calculated fields / Labels that are in the new version of the Epicor base RDD. It is that “fix up” that is pushing the label table to have more than 1024 columns.
There is currently a “Bug” in the RDD import that you may be able to leverage to restore your previous RDD definition - assuming you have or can get an export of it. The bug does not apply the upgrade processing “fix up” logic on the Import. This is not a guarantee that the imported RDD will work as new columns added to the tables defined in the RDD will automatically be included in the Label Table and if a new Calculated Column was added you will get a run time error during the data extraction step in report creation.
From comparison I think we may have been close to the 1024 column limit before and when one new table was added in the RDD to account for Field Service Automation it put us over the limit. It was less time consuming for me to re-create a new RDD that would include any new fields Epicor may have added. This way I have a cleaner RDD going forward.
We are upgrading this weekend and I was having similar problems. In the beginning i tried to exclude the field to get it to work but after 2 or 3 hours I was still in getting errors. In my test environment i just rebuilt the RDD file by copying the system RDD ARFORM and added all the relationships back. I still had to exclude a bunch of field but i got it to work. I am not saying this is the correct way but it worked for me. I only have 3 tables to added and 3 relationship to add back. It took 20 mins.
Jeb
@Rich , do you have any updated info on which minor update of 10.2.300 will have the SQL Wide Table option? I just ran through the .8 release notes and don’t see it yet (unless I missed it).
Thanks!
Fix is in 300.8 and is referenced on the second to last page of the Release Notes. Issue reference is: ERPS-105399 - Having more than 1024 report labels causes an error in the log but it is still possible to run the report
Description in the release notes is a little off-putting as the 1024 column Does Not cause an error and instead the fact the 1024 limit was reached is written to the Server Log.
We were not able to use the SQL “wide table” option so we are instead limiting the number of columns that we will create and populate to 1024. There is still a slight possibility that solution will result in an error from SSRS if you have referenced a Label in the RDL that did not make the 1024 column “cut”. In the unlikely event that you hit that error, you will need to manually Exclude Labels in the Report Data Definition UI - note the UI Toolbar “Action” Options that allow for mass exclusion to help with that.
3 Likes