This discussion is archived
1 Reply Latest reply: Oct 22, 2013 3:14 PM by Sergiusz Wolicki (Oracle) RSS

Datatype CLOB: lossy conversion when migrating from WE8ISO8859P1 to AL32UTF8

vivalavida Journeyer
Currently Being Moderated

We have data of an application in a database of release (characterset WE8ISO8859P1).

The data shall now be migrated into a database of release (characterset AL32UTF8).

When executing the CSSCAN utiltity on the target database ( I get the following result:


Column: INFO
Type  : CLOB
Number of Exceptions         : 116      
Max Post Conversion Data Size: 1489532  

ROWID              Exception Type      Size Cell Data(first 30 bytes)    
------------------ ------------------ ----- ------------------------------
AAA4jBAAKAAACRcAAA lossy conversion 
AAA4jBAAKAAACRcAAB lossy conversion 
AAA4jBAAKAAAvNqAAA lossy conversion 
AAA4jBAAKAABg0nAAC lossy conversion 
AAA4jBAAKAABg0nAAE lossy conversion



From my understanding CLOB objects can store up to (4 gigabytes -1) * (the value of the CHUNK parameter of LOB storage)

of character data. Or is the problem that the characters of some rows can't get converted into AL32UTF8?

Currently I do not understand why "lossy conversion" will happen...


Any help will be appreciated!



  • 1. Re: Datatype CLOB: lossy conversion when migrating from WE8ISO8859P1 to AL32UTF8
    Sergiusz Wolicki (Oracle) Expert
    Currently Being Moderated

    "Lossy conversion" means that some character codes (bytes) in the listed CLOBs are not recognized as valid in WE8ISO8859P1. These codes are usually in the range 0x80-0x9f. You need to take a look at these codes and try to guess what their real meaning is.  In most cases that Oracle customers encounter, these illegal WE8ISO8859P1 codes are in reality WE8MSWIN1252 codes (WE8MSWIN1252 is a binary superset of WE8ISO8859P1). The solution is then to fix the source database character set to the right character set and then export.






  • Correct Answers - 10 points
  • Helpful Answers - 5 points