We are experiencing some issues with bursting compound PDF reports. If the primary sort field value for which we are bursting contain mixed case letters, we are either getting Adobe errors or data mixup among reports. It doesn't happen for all sort values with mixed case which makes it hard to debug. But, if we convert sort values to uppercase, it seems to fix the problem. Has anyone experienced this problem before? We are trying to avoid converting the sort values to uppercase if possible because we use it for display purpose also and there are 100+ reports that gets affected by this change.
Thanks for the help.

This message has been edited. Last edited by: CT,

This could be caused by the fact that in the ASCII collating sequence upper case and lower case letters are not in sequence, Therefor Aa dose not follow or precede AA but comes after AZ.

Thanks. It makes sense. Is there a setting in WebFocus reporting server or somewhere to ignore the case when sorting?
Thanks for the help.

Two approaches to try:

#1. use
TABLE ...
BY myUpperCaseValue NOPRINT
BY mixedCaseValue
-- where myUpperCaseValue is either predefined in the data source table, or DEFINEd using UPCASE() function


#2. perhaps you can hold in desired order (using #1), and then
TABLEF ...
BY mixedCaseValue

- TABLEF will break whenever the sort-key value changes, regardless of direction of the change -- but I'm not sure that Reportcaster bursting will work with TABLEF.

Thanks j.gross and RSquared.
Your responses gave me some clues on what to look for. We are able to find a setting which solved our problem.

COLLATION = enabling keyword
Internal default: CODEPAGE
Establish Binary or Case-Insensitive Collation Sequence. The keyword values are:

CODEPAGE Bases collation sequence on the code page in effect
BINARY Bases the collation sequence on binary values
SRV_CS Bases collation sequence on the LANGUAGE setting, and is case-sensitive
SRV_CI Bases collation sequence on the LANGUAGE setting, and is case-insensitive

We changed this setting from SRV_CI to SRV_CS which fixed our problem
I think when report castor bursts reports, they are in ASCII collating sequence, so the setting SRV_Cl made them out of order with the BY field values in the report.