PA Job Collection fails when collecting Text Analytics for several Fields to Analyze. Customers that have not upgraded to London and create a new Text Analytics Configuration record and use the list "Fields to Analyze" to input several fields may incur into this issue during the collection of the text analytics. PA Data Collection jobs will finish with state "Collected with errors" and in the PA Job Logs you will see a record with level "Error" with a message containing:

Error collecting for 20180529 java.lang.IllegalStateException: Sparse storage format has one field, sys_id
at com.glide.db.meta.SparseStorage.getValue(
at com.glide.schedule.JobExecutor.executeJob(
at com.glide.schedule.JobExecutor.execute(
at com.glide.schedule.GlideScheduleWorker.executeJob(
at com.glide.schedule.GlideScheduleWorker.lambda$process$12(
at com.glide.worker.TransactionalWorkerThread.executeInTransaction(
at com.glide.schedule.GlideScheduleWorker.process(

Steps to Reproduce

On an pre-London instance:

1. Go to Performance Analytics > Text Analytics > Setup Fields.
2. Click on new, make Indicator Source as "Incidents.New", click on the lock icon of the Fields to analyze.
3. Select the fields "Close code", "Close notes", "Description", "Short Description" and click on the lock again to close the list.
4. The list will read "Close code, Close notes, Description, Short description", click on Submit.
5. Go to Performance Analytics > Data Collector > Jobs, click on New.
6. Setup any name, Parameters: Relative, start 1, end 1, both days ago.
7. Run as "System Administrator", Run "On Demand", Collect "Both scores and text index", Save the record.
8. Edit the Indicators related list, add the indicator "Number of New Incidents".
9. Execute the data collection job.

Data collection will fail on both  cores and text indexes with the error message:

"java.lang.IllegalStateException: Sparse storage format has one field, sys_id"


This problem has been fixed in London. The workaround on earlier releases is to increase the max length of the field "Fields to Analyze" for the table Text Index Configurations [pa_text_index_configurations.fields] from "40" to "4000" on the dictionary:

  1. Go to the list of Dictionary Entries by typing in the navigator filter "sys_dictionary.list" or by navigating to System Definition > Dictionary.
  2. Filter the list by "Table" equals to "pa_text_index_configurations" and "Column Name" equals to "fields".
  3. Only one record is listed now in the Dictionary Entries list.
  4. Make sure the column "Max Length" is visible in the list, otherwise, customize the list view and include the column "Max Length".
  5. Double click in the value of the "Max Length" column to make it editable:
  6. Change the value to 4000 and click on the green mark to save it.
  7. Go back to the Performance Analytics > Text Analytics > Setup Fields to fix the Text Index Configuration records.
  8. Edit the fields in "Fields to Analyze" list, this time all the fields will be correctly captured in the list.
  9. Re-run the PA Data Collection to verify that the issue has been fixed.

These steps may be encapsulated in an Update Set to allow to move the workaround across instances. Please test this workaround first in Sub-Production instances (preferably in a recent clone from production) before applying in Production.


Related Problem: PRB1291458

Seen In

There is no data to report.

Fixed In


Associated Community Threads

There is no data to report.

Article Information

Last Updated:2019-06-09 21:40:20